'Weak Dependency Graph [60.0]' ------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} Details: We have computed the following set of weak (innermost) dependency pairs: { i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1)))))))))) , i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1)))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1)) , p^#(s(x1)) -> c_5() , p^#(0(x1)) -> c_6()} The usable rules are: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} The estimated dependency graph contains the following edges: {i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} ==> {p^#(s(x1)) -> c_5()} {i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))))} ==> {p^#(s(x1)) -> c_5()} {j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} ==> {p^#(s(x1)) -> c_5()} {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} ==> {p^#(p(s(x1))) -> c_4(p^#(x1))} {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} ==> {p^#(0(x1)) -> c_6()} {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} ==> {p^#(s(x1)) -> c_5()} {p^#(p(s(x1))) -> c_4(p^#(x1))} ==> {p^#(0(x1)) -> c_6()} {p^#(p(s(x1))) -> c_4(p^#(x1))} ==> {p^#(s(x1)) -> c_5()} {p^#(p(s(x1))) -> c_4(p^#(x1))} ==> {p^#(p(s(x1))) -> c_4(p^#(x1))} We consider the following path(s): 1) { i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , p^#(s(x1)) -> c_5()} The usable rules for this path are the following: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , p^#(s(x1)) -> c_5()} Details: We apply the weight gap principle, strictly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [1] x1 + [1] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [1] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , p^#(s(x1)) -> c_5()} and weakly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , p^#(s(x1)) -> c_5()} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [8] i^#(x1) = [1] x1 + [1] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [2] c_1(x1) = [1] x1 + [1] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))))} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , p^#(s(x1)) -> c_5() , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [8] i^#(x1) = [1] x1 + [9] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [1] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} and weakly orienting the rules { i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , p^#(s(x1)) -> c_5() , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [4] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [1] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [8] i^#(x1) = [1] x1 + [15] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [1] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: {i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , p^#(s(x1)) -> c_5() , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: {i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , p^#(s(x1)) -> c_5() , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_0(21) -> 60 , i_0(21) -> 64 , i_1(87) -> 97 , i_1(87) -> 101 , i_2(127) -> 122 , i_2(127) -> 126 , i_2(138) -> 133 , i_2(138) -> 137 , i_2(208) -> 203 , i_2(208) -> 207 , i_2(219) -> 214 , i_2(219) -> 218 , i_2(289) -> 284 , i_2(289) -> 288 , i_2(300) -> 295 , i_2(300) -> 299 , i_2(370) -> 365 , i_2(370) -> 369 , i_2(381) -> 376 , i_2(381) -> 380 , 0_0(2) -> 2 , 0_0(2) -> 15 , 0_0(2) -> 17 , 0_0(2) -> 19 , 0_0(2) -> 20 , 0_0(2) -> 21 , 0_0(2) -> 22 , 0_0(2) -> 70 , 0_0(2) -> 72 , 0_0(2) -> 74 , 0_0(2) -> 87 , 0_0(2) -> 89 , 0_0(2) -> 381 , 0_0(2) -> 383 , 0_0(2) -> 391 , 0_0(2) -> 393 , 0_0(2) -> 395 , 0_0(4) -> 2 , 0_0(4) -> 15 , 0_0(4) -> 17 , 0_0(4) -> 19 , 0_0(4) -> 20 , 0_0(4) -> 21 , 0_0(4) -> 22 , 0_0(4) -> 70 , 0_0(4) -> 72 , 0_0(4) -> 74 , 0_0(4) -> 87 , 0_0(4) -> 89 , 0_0(4) -> 381 , 0_0(4) -> 383 , 0_0(4) -> 391 , 0_0(4) -> 393 , 0_0(4) -> 395 , 0_0(21) -> 14 , 0_0(21) -> 52 , 0_0(21) -> 56 , 0_0(22) -> 20 , 0_1(82) -> 15 , 0_1(82) -> 17 , 0_1(82) -> 19 , 0_1(82) -> 20 , 0_1(82) -> 21 , 0_1(87) -> 14 , 0_1(87) -> 60 , 0_1(87) -> 64 , 0_1(87) -> 66 , 0_1(87) -> 68 , 0_1(87) -> 69 , 0_1(87) -> 91 , 0_1(87) -> 97 , 0_1(87) -> 101 , 0_1(87) -> 376 , 0_1(87) -> 380 , 0_1(87) -> 389 , 0_2(413) -> 376 , 0_2(413) -> 380 , 0_2(413) -> 389 , p_0(2) -> 15 , p_0(2) -> 17 , p_0(2) -> 19 , p_0(2) -> 20 , p_0(2) -> 21 , p_0(4) -> 15 , p_0(4) -> 17 , p_0(4) -> 19 , p_0(4) -> 20 , p_0(4) -> 21 , p_0(4) -> 22 , p_0(4) -> 70 , p_0(4) -> 72 , p_0(4) -> 74 , p_0(4) -> 87 , p_0(4) -> 89 , p_0(4) -> 381 , p_0(4) -> 383 , p_0(4) -> 391 , p_0(4) -> 393 , p_0(4) -> 395 , p_0(12) -> 11 , p_0(16) -> 15 , p_0(18) -> 15 , p_0(18) -> 17 , p_0(20) -> 15 , p_0(20) -> 17 , p_0(20) -> 19 , p_0(21) -> 20 , p_0(22) -> 15 , p_0(22) -> 17 , p_0(22) -> 19 , p_0(22) -> 20 , p_0(22) -> 21 , p_0(51) -> 14 , p_0(53) -> 14 , p_0(53) -> 52 , p_0(54) -> 53 , p_0(55) -> 14 , p_0(55) -> 52 , p_0(61) -> 60 , p_0(62) -> 61 , p_0(63) -> 60 , p_1(65) -> 60 , p_1(65) -> 64 , p_1(67) -> 14 , p_1(67) -> 60 , p_1(67) -> 64 , p_1(67) -> 66 , p_1(67) -> 69 , p_1(67) -> 91 , p_1(67) -> 97 , p_1(67) -> 101 , p_1(67) -> 376 , p_1(67) -> 380 , p_1(67) -> 389 , p_1(68) -> 376 , p_1(68) -> 380 , p_1(68) -> 389 , p_1(69) -> 389 , p_1(71) -> 70 , p_1(73) -> 70 , p_1(73) -> 72 , p_1(75) -> 70 , p_1(75) -> 72 , p_1(75) -> 74 , p_1(76) -> 75 , p_1(77) -> 76 , p_1(78) -> 77 , p_1(79) -> 76 , p_1(80) -> 75 , p_1(81) -> 70 , p_1(81) -> 72 , p_1(81) -> 74 , p_1(81) -> 87 , p_1(81) -> 89 , p_1(88) -> 87 , p_1(90) -> 14 , p_1(90) -> 69 , p_1(90) -> 97 , p_1(90) -> 101 , p_1(90) -> 376 , p_1(90) -> 380 , p_1(90) -> 389 , p_1(91) -> 376 , p_1(91) -> 380 , p_1(91) -> 389 , p_1(92) -> 14 , p_1(92) -> 69 , p_1(92) -> 91 , p_1(92) -> 97 , p_1(92) -> 101 , p_1(92) -> 376 , p_1(92) -> 380 , p_1(92) -> 389 , p_1(93) -> 92 , p_1(96) -> 97 , p_1(96) -> 101 , p_1(97) -> 376 , p_1(97) -> 380 , p_1(98) -> 97 , p_1(98) -> 101 , p_1(99) -> 98 , p_1(100) -> 97 , p_1(100) -> 101 , p_1(101) -> 376 , p_1(101) -> 380 , p_1(380) -> 389 , p_2(68) -> 389 , p_2(69) -> 389 , p_2(91) -> 389 , p_2(102) -> 97 , p_2(102) -> 101 , p_2(102) -> 376 , p_2(102) -> 380 , p_2(103) -> 376 , p_2(103) -> 380 , p_2(103) -> 389 , p_2(104) -> 97 , p_2(104) -> 101 , p_2(104) -> 103 , p_2(104) -> 376 , p_2(104) -> 380 , p_2(105) -> 376 , p_2(105) -> 380 , p_2(105) -> 389 , p_2(106) -> 389 , p_2(108) -> 107 , p_2(110) -> 107 , p_2(110) -> 109 , p_2(112) -> 107 , p_2(112) -> 109 , p_2(112) -> 111 , p_2(113) -> 112 , p_2(114) -> 113 , p_2(115) -> 114 , p_2(116) -> 113 , p_2(117) -> 112 , p_2(118) -> 107 , p_2(118) -> 109 , p_2(118) -> 111 , p_2(118) -> 127 , p_2(118) -> 129 , p_2(123) -> 122 , p_2(124) -> 123 , p_2(125) -> 122 , p_2(125) -> 126 , p_2(128) -> 127 , p_2(134) -> 133 , p_2(135) -> 134 , p_2(136) -> 133 , p_2(136) -> 137 , p_2(139) -> 138 , p_2(141) -> 138 , p_2(141) -> 140 , p_2(141) -> 148 , p_2(141) -> 150 , p_2(141) -> 152 , p_2(142) -> 122 , p_2(142) -> 126 , p_2(149) -> 148 , p_2(151) -> 148 , p_2(151) -> 150 , p_2(153) -> 148 , p_2(153) -> 150 , p_2(153) -> 152 , p_2(154) -> 153 , p_2(155) -> 154 , p_2(156) -> 155 , p_2(157) -> 154 , p_2(158) -> 153 , p_2(170) -> 133 , p_2(170) -> 137 , p_2(177) -> 176 , p_2(179) -> 176 , p_2(179) -> 178 , p_2(181) -> 176 , p_2(181) -> 178 , p_2(181) -> 180 , p_2(182) -> 181 , p_2(183) -> 182 , p_2(184) -> 183 , p_2(185) -> 182 , p_2(186) -> 181 , p_2(187) -> 176 , p_2(187) -> 178 , p_2(187) -> 180 , p_2(187) -> 208 , p_2(187) -> 210 , p_2(204) -> 203 , p_2(205) -> 204 , p_2(206) -> 203 , p_2(206) -> 207 , p_2(209) -> 208 , p_2(215) -> 214 , p_2(216) -> 215 , p_2(217) -> 214 , p_2(217) -> 218 , p_2(220) -> 219 , p_2(222) -> 219 , p_2(222) -> 221 , p_2(222) -> 229 , p_2(222) -> 231 , p_2(222) -> 233 , p_2(223) -> 203 , p_2(223) -> 207 , p_2(230) -> 229 , p_2(232) -> 229 , p_2(232) -> 231 , p_2(234) -> 229 , p_2(234) -> 231 , p_2(234) -> 233 , p_2(235) -> 234 , p_2(236) -> 235 , p_2(237) -> 236 , p_2(238) -> 235 , p_2(239) -> 234 , p_2(251) -> 214 , p_2(251) -> 218 , p_2(258) -> 257 , p_2(260) -> 257 , p_2(260) -> 259 , p_2(262) -> 257 , p_2(262) -> 259 , p_2(262) -> 261 , p_2(263) -> 262 , p_2(264) -> 263 , p_2(265) -> 264 , p_2(266) -> 263 , p_2(267) -> 262 , p_2(268) -> 257 , p_2(268) -> 259 , p_2(268) -> 261 , p_2(268) -> 289 , p_2(268) -> 291 , p_2(285) -> 284 , p_2(286) -> 285 , p_2(287) -> 284 , p_2(287) -> 288 , p_2(290) -> 289 , p_2(296) -> 295 , p_2(297) -> 296 , p_2(298) -> 295 , p_2(298) -> 299 , p_2(301) -> 300 , p_2(303) -> 300 , p_2(303) -> 302 , p_2(303) -> 310 , p_2(303) -> 312 , p_2(303) -> 314 , p_2(304) -> 284 , p_2(304) -> 288 , p_2(311) -> 310 , p_2(313) -> 310 , p_2(313) -> 312 , p_2(315) -> 310 , p_2(315) -> 312 , p_2(315) -> 314 , p_2(316) -> 315 , p_2(317) -> 316 , p_2(318) -> 317 , p_2(319) -> 316 , p_2(320) -> 315 , p_2(332) -> 295 , p_2(332) -> 299 , p_2(339) -> 338 , p_2(341) -> 338 , p_2(341) -> 340 , p_2(343) -> 338 , p_2(343) -> 340 , p_2(343) -> 342 , p_2(344) -> 343 , p_2(345) -> 344 , p_2(346) -> 345 , p_2(347) -> 344 , p_2(348) -> 343 , p_2(349) -> 338 , p_2(349) -> 340 , p_2(349) -> 342 , p_2(349) -> 370 , p_2(349) -> 372 , p_2(366) -> 365 , p_2(367) -> 366 , p_2(368) -> 365 , p_2(368) -> 369 , p_2(371) -> 370 , p_2(375) -> 376 , p_2(375) -> 380 , p_2(376) -> 389 , p_2(377) -> 376 , p_2(377) -> 380 , p_2(378) -> 377 , p_2(379) -> 376 , p_2(379) -> 380 , p_2(380) -> 389 , p_2(382) -> 381 , p_2(384) -> 381 , p_2(384) -> 383 , p_2(384) -> 391 , p_2(384) -> 393 , p_2(384) -> 395 , p_2(385) -> 365 , p_2(385) -> 369 , p_2(392) -> 391 , p_2(394) -> 391 , p_2(394) -> 393 , p_2(396) -> 391 , p_2(396) -> 393 , p_2(396) -> 395 , p_2(397) -> 396 , p_2(398) -> 397 , p_2(399) -> 398 , p_2(400) -> 397 , p_2(401) -> 396 , s_0(2) -> 4 , s_0(2) -> 15 , s_0(2) -> 17 , s_0(2) -> 19 , s_0(2) -> 20 , s_0(2) -> 21 , s_0(2) -> 22 , s_0(2) -> 70 , s_0(2) -> 72 , s_0(2) -> 74 , s_0(2) -> 87 , s_0(2) -> 89 , s_0(2) -> 381 , s_0(2) -> 383 , s_0(2) -> 391 , s_0(2) -> 393 , s_0(2) -> 395 , s_0(4) -> 4 , s_0(4) -> 15 , s_0(4) -> 17 , s_0(4) -> 19 , s_0(4) -> 20 , s_0(4) -> 21 , s_0(4) -> 22 , s_0(4) -> 70 , s_0(4) -> 72 , s_0(4) -> 74 , s_0(4) -> 87 , s_0(4) -> 89 , s_0(4) -> 381 , s_0(4) -> 383 , s_0(4) -> 391 , s_0(4) -> 393 , s_0(4) -> 395 , s_0(11) -> 10 , s_0(13) -> 12 , s_0(14) -> 11 , s_0(14) -> 13 , s_0(17) -> 16 , s_0(19) -> 18 , s_0(22) -> 2 , s_0(22) -> 15 , s_0(22) -> 17 , s_0(22) -> 19 , s_0(22) -> 20 , s_0(22) -> 21 , s_0(22) -> 22 , s_0(22) -> 70 , s_0(22) -> 72 , s_0(22) -> 74 , s_0(22) -> 87 , s_0(22) -> 89 , s_0(22) -> 381 , s_0(22) -> 383 , s_0(22) -> 391 , s_0(22) -> 393 , s_0(22) -> 395 , s_0(52) -> 51 , s_0(55) -> 54 , s_0(56) -> 53 , s_0(56) -> 55 , s_0(57) -> 14 , s_0(58) -> 57 , s_0(59) -> 58 , s_0(60) -> 59 , s_0(63) -> 62 , s_0(64) -> 61 , s_0(64) -> 63 , s_1(2) -> 75 , s_1(2) -> 81 , s_1(2) -> 338 , s_1(2) -> 340 , s_1(2) -> 342 , s_1(2) -> 370 , s_1(2) -> 372 , s_1(4) -> 75 , s_1(4) -> 81 , s_1(4) -> 338 , s_1(4) -> 340 , s_1(4) -> 342 , s_1(4) -> 370 , s_1(4) -> 372 , s_1(22) -> 75 , s_1(22) -> 81 , s_1(22) -> 338 , s_1(22) -> 340 , s_1(22) -> 342 , s_1(22) -> 370 , s_1(22) -> 372 , s_1(66) -> 65 , s_1(67) -> 93 , s_1(68) -> 67 , s_1(68) -> 92 , s_1(69) -> 14 , s_1(69) -> 60 , s_1(69) -> 64 , s_1(69) -> 66 , s_1(69) -> 68 , s_1(69) -> 69 , s_1(69) -> 91 , s_1(69) -> 97 , s_1(69) -> 101 , s_1(69) -> 376 , s_1(69) -> 380 , s_1(69) -> 389 , s_1(72) -> 71 , s_1(74) -> 73 , s_1(78) -> 85 , s_1(78) -> 176 , s_1(78) -> 178 , s_1(78) -> 180 , s_1(78) -> 208 , s_1(78) -> 210 , s_1(79) -> 78 , s_1(79) -> 219 , s_1(79) -> 221 , s_1(79) -> 229 , s_1(79) -> 231 , s_1(79) -> 233 , s_1(80) -> 77 , s_1(80) -> 79 , s_1(80) -> 257 , s_1(80) -> 259 , s_1(80) -> 261 , s_1(80) -> 289 , s_1(80) -> 291 , s_1(81) -> 76 , s_1(81) -> 80 , s_1(81) -> 300 , s_1(81) -> 302 , s_1(81) -> 310 , s_1(81) -> 312 , s_1(81) -> 314 , s_1(82) -> 75 , s_1(82) -> 81 , s_1(82) -> 338 , s_1(82) -> 340 , s_1(82) -> 342 , s_1(82) -> 370 , s_1(82) -> 372 , s_1(83) -> 70 , s_1(83) -> 72 , s_1(83) -> 74 , s_1(83) -> 82 , s_1(83) -> 87 , s_1(83) -> 89 , s_1(83) -> 381 , s_1(83) -> 383 , s_1(83) -> 391 , s_1(83) -> 393 , s_1(83) -> 395 , s_1(84) -> 83 , s_1(84) -> 107 , s_1(84) -> 109 , s_1(84) -> 111 , s_1(84) -> 127 , s_1(84) -> 129 , s_1(85) -> 84 , s_1(85) -> 138 , s_1(85) -> 140 , s_1(85) -> 148 , s_1(85) -> 150 , s_1(85) -> 152 , s_1(89) -> 88 , s_1(91) -> 90 , s_1(94) -> 14 , s_1(94) -> 69 , s_1(94) -> 376 , s_1(94) -> 380 , s_1(94) -> 389 , s_1(95) -> 94 , s_1(95) -> 389 , s_1(96) -> 95 , s_1(97) -> 96 , s_1(100) -> 99 , s_1(101) -> 98 , s_1(101) -> 100 , s_2(2) -> 384 , s_2(2) -> 396 , s_2(4) -> 384 , s_2(4) -> 396 , s_2(22) -> 384 , s_2(22) -> 396 , s_2(78) -> 222 , s_2(78) -> 234 , s_2(79) -> 262 , s_2(79) -> 268 , s_2(80) -> 303 , s_2(80) -> 315 , s_2(81) -> 343 , s_2(81) -> 349 , s_2(82) -> 384 , s_2(82) -> 396 , s_2(83) -> 112 , s_2(83) -> 118 , s_2(84) -> 141 , s_2(84) -> 153 , s_2(85) -> 181 , s_2(85) -> 187 , s_2(87) -> 420 , s_2(103) -> 102 , s_2(105) -> 104 , s_2(106) -> 97 , s_2(106) -> 101 , s_2(106) -> 103 , s_2(106) -> 105 , s_2(106) -> 376 , s_2(106) -> 380 , s_2(109) -> 108 , s_2(111) -> 110 , s_2(116) -> 115 , s_2(117) -> 114 , s_2(117) -> 116 , s_2(118) -> 113 , s_2(118) -> 117 , s_2(119) -> 69 , s_2(119) -> 376 , s_2(119) -> 380 , s_2(119) -> 389 , s_2(120) -> 119 , s_2(120) -> 389 , s_2(121) -> 120 , s_2(122) -> 121 , s_2(125) -> 124 , s_2(126) -> 123 , s_2(126) -> 125 , s_2(129) -> 128 , s_2(130) -> 106 , s_2(130) -> 376 , s_2(130) -> 380 , s_2(130) -> 389 , s_2(131) -> 130 , s_2(131) -> 389 , s_2(132) -> 131 , s_2(133) -> 132 , s_2(136) -> 135 , s_2(137) -> 134 , s_2(137) -> 136 , s_2(140) -> 139 , s_2(141) -> 154 , s_2(141) -> 158 , s_2(144) -> 142 , s_2(146) -> 122 , s_2(146) -> 126 , s_2(146) -> 144 , s_2(150) -> 149 , s_2(152) -> 151 , s_2(157) -> 156 , s_2(158) -> 155 , s_2(158) -> 157 , s_2(172) -> 170 , s_2(174) -> 133 , s_2(174) -> 137 , s_2(174) -> 172 , s_2(178) -> 177 , s_2(180) -> 179 , s_2(185) -> 184 , s_2(186) -> 183 , s_2(186) -> 185 , s_2(187) -> 182 , s_2(187) -> 186 , s_2(200) -> 146 , s_2(201) -> 200 , s_2(202) -> 201 , s_2(203) -> 202 , s_2(206) -> 205 , s_2(207) -> 204 , s_2(207) -> 206 , s_2(210) -> 209 , s_2(211) -> 174 , s_2(212) -> 211 , s_2(213) -> 212 , s_2(214) -> 213 , s_2(217) -> 216 , s_2(218) -> 215 , s_2(218) -> 217 , s_2(221) -> 220 , s_2(222) -> 235 , s_2(222) -> 239 , s_2(225) -> 223 , s_2(227) -> 203 , s_2(227) -> 207 , s_2(227) -> 225 , s_2(231) -> 230 , s_2(233) -> 232 , s_2(238) -> 237 , s_2(239) -> 236 , s_2(239) -> 238 , s_2(253) -> 251 , s_2(255) -> 214 , s_2(255) -> 218 , s_2(255) -> 253 , s_2(259) -> 258 , s_2(261) -> 260 , s_2(266) -> 265 , s_2(267) -> 264 , s_2(267) -> 266 , s_2(268) -> 263 , s_2(268) -> 267 , s_2(281) -> 227 , s_2(282) -> 281 , s_2(283) -> 282 , s_2(284) -> 283 , s_2(287) -> 286 , s_2(288) -> 285 , s_2(288) -> 287 , s_2(291) -> 290 , s_2(292) -> 255 , s_2(293) -> 292 , s_2(294) -> 293 , s_2(295) -> 294 , s_2(298) -> 297 , s_2(299) -> 296 , s_2(299) -> 298 , s_2(302) -> 301 , s_2(303) -> 316 , s_2(303) -> 320 , s_2(306) -> 304 , s_2(308) -> 284 , s_2(308) -> 288 , s_2(308) -> 306 , s_2(312) -> 311 , s_2(314) -> 313 , s_2(319) -> 318 , s_2(320) -> 317 , s_2(320) -> 319 , s_2(334) -> 332 , s_2(336) -> 295 , s_2(336) -> 299 , s_2(336) -> 334 , s_2(340) -> 339 , s_2(342) -> 341 , s_2(347) -> 346 , s_2(348) -> 345 , s_2(348) -> 347 , s_2(349) -> 344 , s_2(349) -> 348 , s_2(362) -> 308 , s_2(363) -> 362 , s_2(364) -> 363 , s_2(365) -> 364 , s_2(368) -> 367 , s_2(369) -> 366 , s_2(369) -> 368 , s_2(372) -> 371 , s_2(373) -> 336 , s_2(374) -> 373 , s_2(375) -> 374 , s_2(376) -> 375 , s_2(379) -> 378 , s_2(380) -> 377 , s_2(380) -> 379 , s_2(383) -> 382 , s_2(384) -> 397 , s_2(384) -> 401 , s_2(387) -> 385 , s_2(389) -> 365 , s_2(389) -> 369 , s_2(389) -> 387 , s_2(393) -> 392 , s_2(395) -> 394 , s_2(400) -> 399 , s_2(401) -> 398 , s_2(401) -> 400 , s_2(413) -> 420 , s_2(414) -> 413 , s_2(415) -> 414 , s_2(416) -> 415 , s_2(417) -> 416 , s_2(418) -> 417 , s_2(419) -> 418 , s_2(420) -> 419 , j_0(15) -> 14 , j_1(70) -> 69 , j_1(70) -> 376 , j_1(70) -> 380 , j_1(70) -> 389 , j_2(107) -> 106 , j_2(107) -> 376 , j_2(107) -> 380 , j_2(107) -> 389 , j_2(148) -> 146 , j_2(176) -> 174 , j_2(229) -> 227 , j_2(257) -> 255 , j_2(310) -> 308 , j_2(338) -> 336 , j_2(391) -> 389 , i^#_0(2) -> 6 , i^#_0(4) -> 6 , p^#_0(2) -> 8 , p^#_0(4) -> 8 , p^#_0(10) -> 9 , p^#_1(65) -> 86 , c_1_0(9) -> 6 , c_1_1(86) -> 6 , c_5_0() -> 8 , c_5_0() -> 9 , c_5_1() -> 86} 2) {i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))))} The usable rules for this path are the following: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))))} Details: We apply the weight gap principle, strictly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [1] x1 + [1] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [1] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1)))))))))} and weakly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1)))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [8] i^#(x1) = [1] x1 + [1] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [1] x1 + [1] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))))} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [4] i^#(x1) = [1] x1 + [9] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [1] x1 + [1] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} and weakly orienting the rules { i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [4] 0(x1) = [1] x1 + [2] p(x1) = [1] x1 + [1] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [8] i^#(x1) = [1] x1 + [15] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [1] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: {i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: {i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i^#(s(x1)) -> c_1(p^#(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_0(21) -> 60 , i_0(21) -> 64 , i_1(87) -> 97 , i_1(87) -> 101 , i_2(127) -> 122 , i_2(127) -> 126 , i_2(138) -> 133 , i_2(138) -> 137 , i_2(208) -> 203 , i_2(208) -> 207 , i_2(219) -> 214 , i_2(219) -> 218 , i_2(289) -> 284 , i_2(289) -> 288 , i_2(300) -> 295 , i_2(300) -> 299 , i_2(370) -> 365 , i_2(370) -> 369 , i_2(381) -> 376 , i_2(381) -> 380 , 0_0(2) -> 2 , 0_0(2) -> 15 , 0_0(2) -> 17 , 0_0(2) -> 19 , 0_0(2) -> 20 , 0_0(2) -> 21 , 0_0(2) -> 22 , 0_0(2) -> 70 , 0_0(2) -> 72 , 0_0(2) -> 74 , 0_0(2) -> 87 , 0_0(2) -> 89 , 0_0(2) -> 381 , 0_0(2) -> 383 , 0_0(2) -> 391 , 0_0(2) -> 393 , 0_0(2) -> 395 , 0_0(4) -> 2 , 0_0(4) -> 15 , 0_0(4) -> 17 , 0_0(4) -> 19 , 0_0(4) -> 20 , 0_0(4) -> 21 , 0_0(4) -> 22 , 0_0(4) -> 70 , 0_0(4) -> 72 , 0_0(4) -> 74 , 0_0(4) -> 87 , 0_0(4) -> 89 , 0_0(4) -> 381 , 0_0(4) -> 383 , 0_0(4) -> 391 , 0_0(4) -> 393 , 0_0(4) -> 395 , 0_0(21) -> 14 , 0_0(21) -> 52 , 0_0(21) -> 56 , 0_0(22) -> 20 , 0_1(82) -> 15 , 0_1(82) -> 17 , 0_1(82) -> 19 , 0_1(82) -> 20 , 0_1(82) -> 21 , 0_1(87) -> 14 , 0_1(87) -> 60 , 0_1(87) -> 64 , 0_1(87) -> 66 , 0_1(87) -> 68 , 0_1(87) -> 69 , 0_1(87) -> 91 , 0_1(87) -> 97 , 0_1(87) -> 101 , 0_1(87) -> 376 , 0_1(87) -> 380 , 0_1(87) -> 389 , 0_2(413) -> 376 , 0_2(413) -> 380 , 0_2(413) -> 389 , p_0(2) -> 15 , p_0(2) -> 17 , p_0(2) -> 19 , p_0(2) -> 20 , p_0(2) -> 21 , p_0(4) -> 15 , p_0(4) -> 17 , p_0(4) -> 19 , p_0(4) -> 20 , p_0(4) -> 21 , p_0(4) -> 22 , p_0(4) -> 70 , p_0(4) -> 72 , p_0(4) -> 74 , p_0(4) -> 87 , p_0(4) -> 89 , p_0(4) -> 381 , p_0(4) -> 383 , p_0(4) -> 391 , p_0(4) -> 393 , p_0(4) -> 395 , p_0(12) -> 11 , p_0(16) -> 15 , p_0(18) -> 15 , p_0(18) -> 17 , p_0(20) -> 15 , p_0(20) -> 17 , p_0(20) -> 19 , p_0(21) -> 20 , p_0(22) -> 15 , p_0(22) -> 17 , p_0(22) -> 19 , p_0(22) -> 20 , p_0(22) -> 21 , p_0(51) -> 14 , p_0(53) -> 14 , p_0(53) -> 52 , p_0(54) -> 53 , p_0(55) -> 14 , p_0(55) -> 52 , p_0(61) -> 60 , p_0(62) -> 61 , p_0(63) -> 60 , p_1(65) -> 60 , p_1(65) -> 64 , p_1(67) -> 14 , p_1(67) -> 60 , p_1(67) -> 64 , p_1(67) -> 66 , p_1(67) -> 69 , p_1(67) -> 91 , p_1(67) -> 97 , p_1(67) -> 101 , p_1(67) -> 376 , p_1(67) -> 380 , p_1(67) -> 389 , p_1(68) -> 376 , p_1(68) -> 380 , p_1(68) -> 389 , p_1(69) -> 389 , p_1(71) -> 70 , p_1(73) -> 70 , p_1(73) -> 72 , p_1(75) -> 70 , p_1(75) -> 72 , p_1(75) -> 74 , p_1(76) -> 75 , p_1(77) -> 76 , p_1(78) -> 77 , p_1(79) -> 76 , p_1(80) -> 75 , p_1(81) -> 70 , p_1(81) -> 72 , p_1(81) -> 74 , p_1(81) -> 87 , p_1(81) -> 89 , p_1(88) -> 87 , p_1(90) -> 14 , p_1(90) -> 69 , p_1(90) -> 97 , p_1(90) -> 101 , p_1(90) -> 376 , p_1(90) -> 380 , p_1(90) -> 389 , p_1(91) -> 376 , p_1(91) -> 380 , p_1(91) -> 389 , p_1(92) -> 14 , p_1(92) -> 69 , p_1(92) -> 91 , p_1(92) -> 97 , p_1(92) -> 101 , p_1(92) -> 376 , p_1(92) -> 380 , p_1(92) -> 389 , p_1(93) -> 92 , p_1(96) -> 97 , p_1(96) -> 101 , p_1(97) -> 376 , p_1(97) -> 380 , p_1(98) -> 97 , p_1(98) -> 101 , p_1(99) -> 98 , p_1(100) -> 97 , p_1(100) -> 101 , p_1(101) -> 376 , p_1(101) -> 380 , p_1(380) -> 389 , p_2(68) -> 389 , p_2(69) -> 389 , p_2(91) -> 389 , p_2(102) -> 97 , p_2(102) -> 101 , p_2(102) -> 376 , p_2(102) -> 380 , p_2(103) -> 376 , p_2(103) -> 380 , p_2(103) -> 389 , p_2(104) -> 97 , p_2(104) -> 101 , p_2(104) -> 103 , p_2(104) -> 376 , p_2(104) -> 380 , p_2(105) -> 376 , p_2(105) -> 380 , p_2(105) -> 389 , p_2(106) -> 389 , p_2(108) -> 107 , p_2(110) -> 107 , p_2(110) -> 109 , p_2(112) -> 107 , p_2(112) -> 109 , p_2(112) -> 111 , p_2(113) -> 112 , p_2(114) -> 113 , p_2(115) -> 114 , p_2(116) -> 113 , p_2(117) -> 112 , p_2(118) -> 107 , p_2(118) -> 109 , p_2(118) -> 111 , p_2(118) -> 127 , p_2(118) -> 129 , p_2(123) -> 122 , p_2(124) -> 123 , p_2(125) -> 122 , p_2(125) -> 126 , p_2(128) -> 127 , p_2(134) -> 133 , p_2(135) -> 134 , p_2(136) -> 133 , p_2(136) -> 137 , p_2(139) -> 138 , p_2(141) -> 138 , p_2(141) -> 140 , p_2(141) -> 148 , p_2(141) -> 150 , p_2(141) -> 152 , p_2(142) -> 122 , p_2(142) -> 126 , p_2(149) -> 148 , p_2(151) -> 148 , p_2(151) -> 150 , p_2(153) -> 148 , p_2(153) -> 150 , p_2(153) -> 152 , p_2(154) -> 153 , p_2(155) -> 154 , p_2(156) -> 155 , p_2(157) -> 154 , p_2(158) -> 153 , p_2(170) -> 133 , p_2(170) -> 137 , p_2(177) -> 176 , p_2(179) -> 176 , p_2(179) -> 178 , p_2(181) -> 176 , p_2(181) -> 178 , p_2(181) -> 180 , p_2(182) -> 181 , p_2(183) -> 182 , p_2(184) -> 183 , p_2(185) -> 182 , p_2(186) -> 181 , p_2(187) -> 176 , p_2(187) -> 178 , p_2(187) -> 180 , p_2(187) -> 208 , p_2(187) -> 210 , p_2(204) -> 203 , p_2(205) -> 204 , p_2(206) -> 203 , p_2(206) -> 207 , p_2(209) -> 208 , p_2(215) -> 214 , p_2(216) -> 215 , p_2(217) -> 214 , p_2(217) -> 218 , p_2(220) -> 219 , p_2(222) -> 219 , p_2(222) -> 221 , p_2(222) -> 229 , p_2(222) -> 231 , p_2(222) -> 233 , p_2(223) -> 203 , p_2(223) -> 207 , p_2(230) -> 229 , p_2(232) -> 229 , p_2(232) -> 231 , p_2(234) -> 229 , p_2(234) -> 231 , p_2(234) -> 233 , p_2(235) -> 234 , p_2(236) -> 235 , p_2(237) -> 236 , p_2(238) -> 235 , p_2(239) -> 234 , p_2(251) -> 214 , p_2(251) -> 218 , p_2(258) -> 257 , p_2(260) -> 257 , p_2(260) -> 259 , p_2(262) -> 257 , p_2(262) -> 259 , p_2(262) -> 261 , p_2(263) -> 262 , p_2(264) -> 263 , p_2(265) -> 264 , p_2(266) -> 263 , p_2(267) -> 262 , p_2(268) -> 257 , p_2(268) -> 259 , p_2(268) -> 261 , p_2(268) -> 289 , p_2(268) -> 291 , p_2(285) -> 284 , p_2(286) -> 285 , p_2(287) -> 284 , p_2(287) -> 288 , p_2(290) -> 289 , p_2(296) -> 295 , p_2(297) -> 296 , p_2(298) -> 295 , p_2(298) -> 299 , p_2(301) -> 300 , p_2(303) -> 300 , p_2(303) -> 302 , p_2(303) -> 310 , p_2(303) -> 312 , p_2(303) -> 314 , p_2(304) -> 284 , p_2(304) -> 288 , p_2(311) -> 310 , p_2(313) -> 310 , p_2(313) -> 312 , p_2(315) -> 310 , p_2(315) -> 312 , p_2(315) -> 314 , p_2(316) -> 315 , p_2(317) -> 316 , p_2(318) -> 317 , p_2(319) -> 316 , p_2(320) -> 315 , p_2(332) -> 295 , p_2(332) -> 299 , p_2(339) -> 338 , p_2(341) -> 338 , p_2(341) -> 340 , p_2(343) -> 338 , p_2(343) -> 340 , p_2(343) -> 342 , p_2(344) -> 343 , p_2(345) -> 344 , p_2(346) -> 345 , p_2(347) -> 344 , p_2(348) -> 343 , p_2(349) -> 338 , p_2(349) -> 340 , p_2(349) -> 342 , p_2(349) -> 370 , p_2(349) -> 372 , p_2(366) -> 365 , p_2(367) -> 366 , p_2(368) -> 365 , p_2(368) -> 369 , p_2(371) -> 370 , p_2(375) -> 376 , p_2(375) -> 380 , p_2(376) -> 389 , p_2(377) -> 376 , p_2(377) -> 380 , p_2(378) -> 377 , p_2(379) -> 376 , p_2(379) -> 380 , p_2(380) -> 389 , p_2(382) -> 381 , p_2(384) -> 381 , p_2(384) -> 383 , p_2(384) -> 391 , p_2(384) -> 393 , p_2(384) -> 395 , p_2(385) -> 365 , p_2(385) -> 369 , p_2(392) -> 391 , p_2(394) -> 391 , p_2(394) -> 393 , p_2(396) -> 391 , p_2(396) -> 393 , p_2(396) -> 395 , p_2(397) -> 396 , p_2(398) -> 397 , p_2(399) -> 398 , p_2(400) -> 397 , p_2(401) -> 396 , s_0(2) -> 4 , s_0(2) -> 15 , s_0(2) -> 17 , s_0(2) -> 19 , s_0(2) -> 20 , s_0(2) -> 21 , s_0(2) -> 22 , s_0(2) -> 70 , s_0(2) -> 72 , s_0(2) -> 74 , s_0(2) -> 87 , s_0(2) -> 89 , s_0(2) -> 381 , s_0(2) -> 383 , s_0(2) -> 391 , s_0(2) -> 393 , s_0(2) -> 395 , s_0(4) -> 4 , s_0(4) -> 15 , s_0(4) -> 17 , s_0(4) -> 19 , s_0(4) -> 20 , s_0(4) -> 21 , s_0(4) -> 22 , s_0(4) -> 70 , s_0(4) -> 72 , s_0(4) -> 74 , s_0(4) -> 87 , s_0(4) -> 89 , s_0(4) -> 381 , s_0(4) -> 383 , s_0(4) -> 391 , s_0(4) -> 393 , s_0(4) -> 395 , s_0(11) -> 10 , s_0(13) -> 12 , s_0(14) -> 11 , s_0(14) -> 13 , s_0(17) -> 16 , s_0(19) -> 18 , s_0(22) -> 2 , s_0(22) -> 15 , s_0(22) -> 17 , s_0(22) -> 19 , s_0(22) -> 20 , s_0(22) -> 21 , s_0(22) -> 22 , s_0(22) -> 70 , s_0(22) -> 72 , s_0(22) -> 74 , s_0(22) -> 87 , s_0(22) -> 89 , s_0(22) -> 381 , s_0(22) -> 383 , s_0(22) -> 391 , s_0(22) -> 393 , s_0(22) -> 395 , s_0(52) -> 51 , s_0(55) -> 54 , s_0(56) -> 53 , s_0(56) -> 55 , s_0(57) -> 14 , s_0(58) -> 57 , s_0(59) -> 58 , s_0(60) -> 59 , s_0(63) -> 62 , s_0(64) -> 61 , s_0(64) -> 63 , s_1(2) -> 75 , s_1(2) -> 81 , s_1(2) -> 338 , s_1(2) -> 340 , s_1(2) -> 342 , s_1(2) -> 370 , s_1(2) -> 372 , s_1(4) -> 75 , s_1(4) -> 81 , s_1(4) -> 338 , s_1(4) -> 340 , s_1(4) -> 342 , s_1(4) -> 370 , s_1(4) -> 372 , s_1(22) -> 75 , s_1(22) -> 81 , s_1(22) -> 338 , s_1(22) -> 340 , s_1(22) -> 342 , s_1(22) -> 370 , s_1(22) -> 372 , s_1(66) -> 65 , s_1(67) -> 93 , s_1(68) -> 67 , s_1(68) -> 92 , s_1(69) -> 14 , s_1(69) -> 60 , s_1(69) -> 64 , s_1(69) -> 66 , s_1(69) -> 68 , s_1(69) -> 69 , s_1(69) -> 91 , s_1(69) -> 97 , s_1(69) -> 101 , s_1(69) -> 376 , s_1(69) -> 380 , s_1(69) -> 389 , s_1(72) -> 71 , s_1(74) -> 73 , s_1(78) -> 85 , s_1(78) -> 176 , s_1(78) -> 178 , s_1(78) -> 180 , s_1(78) -> 208 , s_1(78) -> 210 , s_1(79) -> 78 , s_1(79) -> 219 , s_1(79) -> 221 , s_1(79) -> 229 , s_1(79) -> 231 , s_1(79) -> 233 , s_1(80) -> 77 , s_1(80) -> 79 , s_1(80) -> 257 , s_1(80) -> 259 , s_1(80) -> 261 , s_1(80) -> 289 , s_1(80) -> 291 , s_1(81) -> 76 , s_1(81) -> 80 , s_1(81) -> 300 , s_1(81) -> 302 , s_1(81) -> 310 , s_1(81) -> 312 , s_1(81) -> 314 , s_1(82) -> 75 , s_1(82) -> 81 , s_1(82) -> 338 , s_1(82) -> 340 , s_1(82) -> 342 , s_1(82) -> 370 , s_1(82) -> 372 , s_1(83) -> 70 , s_1(83) -> 72 , s_1(83) -> 74 , s_1(83) -> 82 , s_1(83) -> 87 , s_1(83) -> 89 , s_1(83) -> 381 , s_1(83) -> 383 , s_1(83) -> 391 , s_1(83) -> 393 , s_1(83) -> 395 , s_1(84) -> 83 , s_1(84) -> 107 , s_1(84) -> 109 , s_1(84) -> 111 , s_1(84) -> 127 , s_1(84) -> 129 , s_1(85) -> 84 , s_1(85) -> 138 , s_1(85) -> 140 , s_1(85) -> 148 , s_1(85) -> 150 , s_1(85) -> 152 , s_1(89) -> 88 , s_1(91) -> 90 , s_1(94) -> 14 , s_1(94) -> 69 , s_1(94) -> 376 , s_1(94) -> 380 , s_1(94) -> 389 , s_1(95) -> 94 , s_1(95) -> 389 , s_1(96) -> 95 , s_1(97) -> 96 , s_1(100) -> 99 , s_1(101) -> 98 , s_1(101) -> 100 , s_2(2) -> 384 , s_2(2) -> 396 , s_2(4) -> 384 , s_2(4) -> 396 , s_2(22) -> 384 , s_2(22) -> 396 , s_2(78) -> 222 , s_2(78) -> 234 , s_2(79) -> 262 , s_2(79) -> 268 , s_2(80) -> 303 , s_2(80) -> 315 , s_2(81) -> 343 , s_2(81) -> 349 , s_2(82) -> 384 , s_2(82) -> 396 , s_2(83) -> 112 , s_2(83) -> 118 , s_2(84) -> 141 , s_2(84) -> 153 , s_2(85) -> 181 , s_2(85) -> 187 , s_2(87) -> 420 , s_2(103) -> 102 , s_2(105) -> 104 , s_2(106) -> 97 , s_2(106) -> 101 , s_2(106) -> 103 , s_2(106) -> 105 , s_2(106) -> 376 , s_2(106) -> 380 , s_2(109) -> 108 , s_2(111) -> 110 , s_2(116) -> 115 , s_2(117) -> 114 , s_2(117) -> 116 , s_2(118) -> 113 , s_2(118) -> 117 , s_2(119) -> 69 , s_2(119) -> 376 , s_2(119) -> 380 , s_2(119) -> 389 , s_2(120) -> 119 , s_2(120) -> 389 , s_2(121) -> 120 , s_2(122) -> 121 , s_2(125) -> 124 , s_2(126) -> 123 , s_2(126) -> 125 , s_2(129) -> 128 , s_2(130) -> 106 , s_2(130) -> 376 , s_2(130) -> 380 , s_2(130) -> 389 , s_2(131) -> 130 , s_2(131) -> 389 , s_2(132) -> 131 , s_2(133) -> 132 , s_2(136) -> 135 , s_2(137) -> 134 , s_2(137) -> 136 , s_2(140) -> 139 , s_2(141) -> 154 , s_2(141) -> 158 , s_2(144) -> 142 , s_2(146) -> 122 , s_2(146) -> 126 , s_2(146) -> 144 , s_2(150) -> 149 , s_2(152) -> 151 , s_2(157) -> 156 , s_2(158) -> 155 , s_2(158) -> 157 , s_2(172) -> 170 , s_2(174) -> 133 , s_2(174) -> 137 , s_2(174) -> 172 , s_2(178) -> 177 , s_2(180) -> 179 , s_2(185) -> 184 , s_2(186) -> 183 , s_2(186) -> 185 , s_2(187) -> 182 , s_2(187) -> 186 , s_2(200) -> 146 , s_2(201) -> 200 , s_2(202) -> 201 , s_2(203) -> 202 , s_2(206) -> 205 , s_2(207) -> 204 , s_2(207) -> 206 , s_2(210) -> 209 , s_2(211) -> 174 , s_2(212) -> 211 , s_2(213) -> 212 , s_2(214) -> 213 , s_2(217) -> 216 , s_2(218) -> 215 , s_2(218) -> 217 , s_2(221) -> 220 , s_2(222) -> 235 , s_2(222) -> 239 , s_2(225) -> 223 , s_2(227) -> 203 , s_2(227) -> 207 , s_2(227) -> 225 , s_2(231) -> 230 , s_2(233) -> 232 , s_2(238) -> 237 , s_2(239) -> 236 , s_2(239) -> 238 , s_2(253) -> 251 , s_2(255) -> 214 , s_2(255) -> 218 , s_2(255) -> 253 , s_2(259) -> 258 , s_2(261) -> 260 , s_2(266) -> 265 , s_2(267) -> 264 , s_2(267) -> 266 , s_2(268) -> 263 , s_2(268) -> 267 , s_2(281) -> 227 , s_2(282) -> 281 , s_2(283) -> 282 , s_2(284) -> 283 , s_2(287) -> 286 , s_2(288) -> 285 , s_2(288) -> 287 , s_2(291) -> 290 , s_2(292) -> 255 , s_2(293) -> 292 , s_2(294) -> 293 , s_2(295) -> 294 , s_2(298) -> 297 , s_2(299) -> 296 , s_2(299) -> 298 , s_2(302) -> 301 , s_2(303) -> 316 , s_2(303) -> 320 , s_2(306) -> 304 , s_2(308) -> 284 , s_2(308) -> 288 , s_2(308) -> 306 , s_2(312) -> 311 , s_2(314) -> 313 , s_2(319) -> 318 , s_2(320) -> 317 , s_2(320) -> 319 , s_2(334) -> 332 , s_2(336) -> 295 , s_2(336) -> 299 , s_2(336) -> 334 , s_2(340) -> 339 , s_2(342) -> 341 , s_2(347) -> 346 , s_2(348) -> 345 , s_2(348) -> 347 , s_2(349) -> 344 , s_2(349) -> 348 , s_2(362) -> 308 , s_2(363) -> 362 , s_2(364) -> 363 , s_2(365) -> 364 , s_2(368) -> 367 , s_2(369) -> 366 , s_2(369) -> 368 , s_2(372) -> 371 , s_2(373) -> 336 , s_2(374) -> 373 , s_2(375) -> 374 , s_2(376) -> 375 , s_2(379) -> 378 , s_2(380) -> 377 , s_2(380) -> 379 , s_2(383) -> 382 , s_2(384) -> 397 , s_2(384) -> 401 , s_2(387) -> 385 , s_2(389) -> 365 , s_2(389) -> 369 , s_2(389) -> 387 , s_2(393) -> 392 , s_2(395) -> 394 , s_2(400) -> 399 , s_2(401) -> 398 , s_2(401) -> 400 , s_2(413) -> 420 , s_2(414) -> 413 , s_2(415) -> 414 , s_2(416) -> 415 , s_2(417) -> 416 , s_2(418) -> 417 , s_2(419) -> 418 , s_2(420) -> 419 , j_0(15) -> 14 , j_1(70) -> 69 , j_1(70) -> 376 , j_1(70) -> 380 , j_1(70) -> 389 , j_2(107) -> 106 , j_2(107) -> 376 , j_2(107) -> 380 , j_2(107) -> 389 , j_2(148) -> 146 , j_2(176) -> 174 , j_2(229) -> 227 , j_2(257) -> 255 , j_2(310) -> 308 , j_2(338) -> 336 , j_2(391) -> 389 , i^#_0(2) -> 6 , i^#_0(4) -> 6 , p^#_0(2) -> 8 , p^#_0(4) -> 8 , p^#_0(10) -> 9 , p^#_1(65) -> 86 , c_1_0(9) -> 6 , c_1_1(86) -> 6} 3) { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(s(x1)) -> c_5()} The usable rules for this path are the following: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(s(x1)) -> c_5()} Details: We apply the weight gap principle, strictly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {p^#(s(x1)) -> c_5()} and weakly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {p^#(s(x1)) -> c_5()} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [8] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} and weakly orienting the rules { p^#(s(x1)) -> c_5() , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [9] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(s(x1)) -> c_5() , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(s(x1)) -> c_5() , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_1(8) -> 7 , i_1(8) -> 16 , i_1(8) -> 31 , 0_0(2) -> 2 , 0_0(2) -> 8 , 0_0(2) -> 10 , 0_0(2) -> 17 , 0_0(2) -> 19 , 0_0(2) -> 21 , 0_1(8) -> 7 , 0_1(8) -> 13 , 0_1(8) -> 15 , 0_1(8) -> 16 , 0_1(8) -> 31 , p_1(4) -> 16 , p_1(4) -> 31 , p_1(5) -> 4 , p_1(9) -> 8 , p_1(11) -> 8 , p_1(11) -> 10 , p_1(12) -> 7 , p_1(12) -> 16 , p_1(12) -> 31 , p_1(14) -> 7 , p_1(14) -> 13 , p_1(14) -> 16 , p_1(14) -> 31 , p_1(18) -> 17 , p_1(20) -> 17 , p_1(20) -> 19 , p_1(22) -> 17 , p_1(22) -> 19 , p_1(22) -> 21 , p_1(23) -> 22 , p_1(24) -> 23 , p_1(25) -> 24 , p_1(30) -> 16 , p_2(6) -> 16 , p_2(6) -> 31 , p_2(11) -> 17 , p_2(11) -> 19 , p_2(11) -> 21 , p_2(26) -> 23 , p_2(27) -> 22 , s_0(2) -> 2 , s_0(2) -> 8 , s_0(2) -> 10 , s_0(2) -> 17 , s_0(2) -> 19 , s_0(2) -> 21 , s_1(2) -> 11 , s_1(2) -> 22 , s_1(6) -> 5 , s_1(7) -> 4 , s_1(7) -> 6 , s_1(10) -> 9 , s_1(11) -> 23 , s_1(11) -> 27 , s_1(13) -> 12 , s_1(15) -> 14 , s_1(16) -> 7 , s_1(16) -> 13 , s_1(16) -> 15 , s_1(16) -> 16 , s_1(16) -> 31 , s_1(19) -> 18 , s_1(21) -> 20 , s_1(26) -> 25 , s_1(27) -> 24 , s_1(27) -> 26 , s_1(28) -> 16 , s_1(29) -> 28 , s_1(30) -> 29 , s_1(31) -> 30 , j_1(17) -> 16 , p^#_0(2) -> 1 , p^#_1(4) -> 3 , j^#_0(2) -> 1 , c_3_1(3) -> 1 , c_5_0() -> 1 , c_5_1() -> 3} 4) { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1)) , p^#(s(x1)) -> c_5()} The usable rules for this path are the following: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1)) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(s(x1)) -> c_5()} Details: We apply the weight gap principle, strictly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(s(x1)) -> c_5()} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(s(x1)) -> c_5()} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [1] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(s(x1)) -> c_5()} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [1] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} and weakly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(s(x1)) -> c_5()} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [13] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [3] c_4(x1) = [1] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(s(x1)) -> c_5()} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(s(x1)) -> c_5()} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_1(8) -> 7 , i_1(8) -> 17 , i_1(8) -> 32 , 0_0(2) -> 2 , 0_0(2) -> 8 , 0_0(2) -> 10 , 0_0(2) -> 18 , 0_0(2) -> 20 , 0_0(2) -> 22 , 0_1(8) -> 7 , 0_1(8) -> 14 , 0_1(8) -> 16 , 0_1(8) -> 17 , 0_1(8) -> 32 , p_1(4) -> 17 , p_1(4) -> 32 , p_1(5) -> 4 , p_1(9) -> 8 , p_1(11) -> 8 , p_1(11) -> 10 , p_1(13) -> 7 , p_1(13) -> 17 , p_1(13) -> 32 , p_1(15) -> 7 , p_1(15) -> 14 , p_1(15) -> 17 , p_1(15) -> 32 , p_1(19) -> 18 , p_1(21) -> 18 , p_1(21) -> 20 , p_1(23) -> 18 , p_1(23) -> 20 , p_1(23) -> 22 , p_1(24) -> 23 , p_1(25) -> 24 , p_1(26) -> 25 , p_1(31) -> 17 , p_2(6) -> 17 , p_2(6) -> 32 , p_2(11) -> 18 , p_2(11) -> 20 , p_2(11) -> 22 , p_2(27) -> 24 , p_2(28) -> 23 , s_0(2) -> 2 , s_0(2) -> 8 , s_0(2) -> 10 , s_0(2) -> 18 , s_0(2) -> 20 , s_0(2) -> 22 , s_1(2) -> 11 , s_1(2) -> 23 , s_1(6) -> 5 , s_1(7) -> 4 , s_1(7) -> 6 , s_1(10) -> 9 , s_1(11) -> 24 , s_1(11) -> 28 , s_1(14) -> 13 , s_1(16) -> 15 , s_1(17) -> 7 , s_1(17) -> 14 , s_1(17) -> 16 , s_1(17) -> 17 , s_1(17) -> 32 , s_1(20) -> 19 , s_1(22) -> 21 , s_1(27) -> 26 , s_1(28) -> 25 , s_1(28) -> 27 , s_1(29) -> 17 , s_1(30) -> 29 , s_1(31) -> 30 , s_1(32) -> 31 , j_1(18) -> 17 , p^#_0(2) -> 1 , p^#_1(4) -> 3 , p^#_2(6) -> 12 , j^#_0(2) -> 1 , c_3_1(3) -> 1 , c_4_2(12) -> 3 , c_5_0() -> 1 , c_5_1() -> 3 , c_5_2() -> 12} 5) { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1)) , p^#(0(x1)) -> c_6()} The usable rules for this path are the following: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1)) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(0(x1)) -> c_6()} Details: We apply the weight gap principle, strictly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(0(x1)) -> c_6()} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(0(x1)) -> c_6()} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [1] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(0(x1)) -> c_6()} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [1] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} and weakly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(0(x1)) -> c_6()} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [13] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [3] c_4(x1) = [1] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(0(x1)) -> c_6()} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p^#(0(x1)) -> c_6()} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_1(8) -> 7 , i_1(8) -> 17 , i_1(8) -> 32 , 0_0(2) -> 2 , 0_0(2) -> 8 , 0_0(2) -> 10 , 0_0(2) -> 18 , 0_0(2) -> 20 , 0_0(2) -> 22 , 0_1(8) -> 7 , 0_1(8) -> 14 , 0_1(8) -> 16 , 0_1(8) -> 17 , 0_1(8) -> 32 , p_1(4) -> 17 , p_1(4) -> 32 , p_1(5) -> 4 , p_1(9) -> 8 , p_1(11) -> 8 , p_1(11) -> 10 , p_1(13) -> 7 , p_1(13) -> 17 , p_1(13) -> 32 , p_1(15) -> 7 , p_1(15) -> 14 , p_1(15) -> 17 , p_1(15) -> 32 , p_1(19) -> 18 , p_1(21) -> 18 , p_1(21) -> 20 , p_1(23) -> 18 , p_1(23) -> 20 , p_1(23) -> 22 , p_1(24) -> 23 , p_1(25) -> 24 , p_1(26) -> 25 , p_1(31) -> 17 , p_2(6) -> 17 , p_2(6) -> 32 , p_2(11) -> 18 , p_2(11) -> 20 , p_2(11) -> 22 , p_2(27) -> 24 , p_2(28) -> 23 , s_0(2) -> 2 , s_0(2) -> 8 , s_0(2) -> 10 , s_0(2) -> 18 , s_0(2) -> 20 , s_0(2) -> 22 , s_1(2) -> 11 , s_1(2) -> 23 , s_1(6) -> 5 , s_1(7) -> 4 , s_1(7) -> 6 , s_1(10) -> 9 , s_1(11) -> 24 , s_1(11) -> 28 , s_1(14) -> 13 , s_1(16) -> 15 , s_1(17) -> 7 , s_1(17) -> 14 , s_1(17) -> 16 , s_1(17) -> 17 , s_1(17) -> 32 , s_1(20) -> 19 , s_1(22) -> 21 , s_1(27) -> 26 , s_1(28) -> 25 , s_1(28) -> 27 , s_1(29) -> 17 , s_1(30) -> 29 , s_1(31) -> 30 , s_1(32) -> 31 , j_1(18) -> 17 , p^#_0(2) -> 1 , p^#_1(4) -> 3 , p^#_2(6) -> 12 , j^#_0(2) -> 1 , c_3_1(3) -> 1 , c_4_2(12) -> 3 , c_6_0() -> 1} 6) {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} The usable rules for this path are the following: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: We apply the weight gap principle, strictly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} and weakly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [9] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_1(8) -> 7 , i_1(8) -> 16 , i_1(8) -> 31 , 0_0(2) -> 2 , 0_0(2) -> 8 , 0_0(2) -> 10 , 0_0(2) -> 17 , 0_0(2) -> 19 , 0_0(2) -> 21 , 0_1(8) -> 7 , 0_1(8) -> 13 , 0_1(8) -> 15 , 0_1(8) -> 16 , 0_1(8) -> 31 , p_1(4) -> 16 , p_1(4) -> 31 , p_1(5) -> 4 , p_1(9) -> 8 , p_1(11) -> 8 , p_1(11) -> 10 , p_1(12) -> 7 , p_1(12) -> 16 , p_1(12) -> 31 , p_1(14) -> 7 , p_1(14) -> 13 , p_1(14) -> 16 , p_1(14) -> 31 , p_1(18) -> 17 , p_1(20) -> 17 , p_1(20) -> 19 , p_1(22) -> 17 , p_1(22) -> 19 , p_1(22) -> 21 , p_1(23) -> 22 , p_1(24) -> 23 , p_1(25) -> 24 , p_1(30) -> 16 , p_2(6) -> 16 , p_2(6) -> 31 , p_2(11) -> 17 , p_2(11) -> 19 , p_2(11) -> 21 , p_2(26) -> 23 , p_2(27) -> 22 , s_0(2) -> 2 , s_0(2) -> 8 , s_0(2) -> 10 , s_0(2) -> 17 , s_0(2) -> 19 , s_0(2) -> 21 , s_1(2) -> 11 , s_1(2) -> 22 , s_1(6) -> 5 , s_1(7) -> 4 , s_1(7) -> 6 , s_1(10) -> 9 , s_1(11) -> 23 , s_1(11) -> 27 , s_1(13) -> 12 , s_1(15) -> 14 , s_1(16) -> 7 , s_1(16) -> 13 , s_1(16) -> 15 , s_1(16) -> 16 , s_1(16) -> 31 , s_1(19) -> 18 , s_1(21) -> 20 , s_1(26) -> 25 , s_1(27) -> 24 , s_1(27) -> 26 , s_1(28) -> 16 , s_1(29) -> 28 , s_1(30) -> 29 , s_1(31) -> 30 , j_1(17) -> 16 , p^#_0(2) -> 1 , p^#_1(4) -> 3 , j^#_0(2) -> 1 , c_3_1(3) -> 1} 7) { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} The usable rules for this path are the following: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} Details: We apply the weight gap principle, strictly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [1] x1 + [1] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1)))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [4] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [1] c_4(x1) = [1] x1 + [1] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} and weakly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [9] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [1] x1 + [3] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , p^#(p(s(x1))) -> c_4(p^#(x1))} Weak Rules: { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1))))))))))))))))))} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_1(8) -> 7 , i_1(8) -> 17 , i_1(8) -> 32 , 0_0(2) -> 2 , 0_0(2) -> 8 , 0_0(2) -> 10 , 0_0(2) -> 18 , 0_0(2) -> 20 , 0_0(2) -> 22 , 0_1(8) -> 7 , 0_1(8) -> 14 , 0_1(8) -> 16 , 0_1(8) -> 17 , 0_1(8) -> 32 , p_1(4) -> 17 , p_1(4) -> 32 , p_1(5) -> 4 , p_1(9) -> 8 , p_1(11) -> 8 , p_1(11) -> 10 , p_1(13) -> 7 , p_1(13) -> 17 , p_1(13) -> 32 , p_1(15) -> 7 , p_1(15) -> 14 , p_1(15) -> 17 , p_1(15) -> 32 , p_1(19) -> 18 , p_1(21) -> 18 , p_1(21) -> 20 , p_1(23) -> 18 , p_1(23) -> 20 , p_1(23) -> 22 , p_1(24) -> 23 , p_1(25) -> 24 , p_1(26) -> 25 , p_1(31) -> 17 , p_2(6) -> 17 , p_2(6) -> 32 , p_2(11) -> 18 , p_2(11) -> 20 , p_2(11) -> 22 , p_2(27) -> 24 , p_2(28) -> 23 , s_0(2) -> 2 , s_0(2) -> 8 , s_0(2) -> 10 , s_0(2) -> 18 , s_0(2) -> 20 , s_0(2) -> 22 , s_1(2) -> 11 , s_1(2) -> 23 , s_1(6) -> 5 , s_1(7) -> 4 , s_1(7) -> 6 , s_1(10) -> 9 , s_1(11) -> 24 , s_1(11) -> 28 , s_1(14) -> 13 , s_1(16) -> 15 , s_1(17) -> 7 , s_1(17) -> 14 , s_1(17) -> 16 , s_1(17) -> 17 , s_1(17) -> 32 , s_1(20) -> 19 , s_1(22) -> 21 , s_1(27) -> 26 , s_1(28) -> 25 , s_1(28) -> 27 , s_1(29) -> 17 , s_1(30) -> 29 , s_1(31) -> 30 , s_1(32) -> 31 , j_1(18) -> 17 , p^#_0(2) -> 1 , p^#_1(4) -> 3 , p^#_2(6) -> 12 , j^#_0(2) -> 1 , c_3_1(3) -> 1 , c_4_2(12) -> 3} 8) { j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(0(x1)) -> c_6()} The usable rules for this path are the following: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} We have applied the subprocessor on the union of usable rules and weak (innermost) dependency pairs. 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost runtime-complexity with respect to Rules: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1)))))))))) , p^#(0(x1)) -> c_6()} Details: We apply the weight gap principle, strictly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} and weakly orienting the rules {} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: Interpretation Functions: i(x1) = [1] x1 + [1] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [9] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor We apply the weight gap principle, strictly orienting the rules { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , p^#(0(x1)) -> c_6()} and weakly orienting the rules { i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , p^#(0(x1)) -> c_6()} Details: Interpretation Functions: i(x1) = [1] x1 + [8] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [1] x1 + [1] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [9] c_2(x1) = [0] x1 + [0] c_3(x1) = [1] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'fastest of 'combine', 'Bounds with default enrichment', 'Bounds with default enrichment'' ------------------------------------------------------------------------------------------ Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Weak Rules: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , p^#(0(x1)) -> c_6() , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: The problem was solved by processor 'Bounds with default enrichment': 'Bounds with default enrichment' -------------------------------- Answer: YES(?,O(n^1)) Input Problem: innermost relative runtime-complexity with respect to Strict Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j(s(x1)) -> s(s(s(s(p(p(s(s(i(p(s(p(s(x1)))))))))))))} Weak Rules: { j(0(x1)) -> p(s(p(p(s(s(0(p(s(p(s(x1))))))))))) , p^#(0(x1)) -> c_6() , i(0(x1)) -> p(s(p(s(0(p(s(p(s(x1))))))))) , i(s(x1)) -> p(s(p(s(s(j(p(s(p(s(p(p(p(p(s(s(s(s(x1)))))))))))))))))) , j^#(s(x1)) -> c_3(p^#(p(s(s(i(p(s(p(s(x1))))))))))} Details: The problem is Match-bounded by 2. The enriched problem is compatible with the following automaton: { i_1(8) -> 7 , i_1(8) -> 16 , i_1(8) -> 31 , 0_0(2) -> 2 , 0_0(2) -> 8 , 0_0(2) -> 10 , 0_0(2) -> 17 , 0_0(2) -> 19 , 0_0(2) -> 21 , 0_1(8) -> 7 , 0_1(8) -> 13 , 0_1(8) -> 15 , 0_1(8) -> 16 , 0_1(8) -> 31 , p_1(4) -> 16 , p_1(4) -> 31 , p_1(5) -> 4 , p_1(9) -> 8 , p_1(11) -> 8 , p_1(11) -> 10 , p_1(12) -> 7 , p_1(12) -> 16 , p_1(12) -> 31 , p_1(14) -> 7 , p_1(14) -> 13 , p_1(14) -> 16 , p_1(14) -> 31 , p_1(18) -> 17 , p_1(20) -> 17 , p_1(20) -> 19 , p_1(22) -> 17 , p_1(22) -> 19 , p_1(22) -> 21 , p_1(23) -> 22 , p_1(24) -> 23 , p_1(25) -> 24 , p_1(30) -> 16 , p_2(6) -> 16 , p_2(6) -> 31 , p_2(11) -> 17 , p_2(11) -> 19 , p_2(11) -> 21 , p_2(26) -> 23 , p_2(27) -> 22 , s_0(2) -> 2 , s_0(2) -> 8 , s_0(2) -> 10 , s_0(2) -> 17 , s_0(2) -> 19 , s_0(2) -> 21 , s_1(2) -> 11 , s_1(2) -> 22 , s_1(6) -> 5 , s_1(7) -> 4 , s_1(7) -> 6 , s_1(10) -> 9 , s_1(11) -> 23 , s_1(11) -> 27 , s_1(13) -> 12 , s_1(15) -> 14 , s_1(16) -> 7 , s_1(16) -> 13 , s_1(16) -> 15 , s_1(16) -> 16 , s_1(16) -> 31 , s_1(19) -> 18 , s_1(21) -> 20 , s_1(26) -> 25 , s_1(27) -> 24 , s_1(27) -> 26 , s_1(28) -> 16 , s_1(29) -> 28 , s_1(30) -> 29 , s_1(31) -> 30 , j_1(17) -> 16 , p^#_0(2) -> 1 , p^#_1(4) -> 3 , j^#_0(2) -> 1 , c_3_1(3) -> 1 , c_6_0() -> 1} 9) { j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1)))))))))))) , p^#(s(x1)) -> c_5()} The usable rules for this path are the following: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} We have oriented the usable rules with the following strongly linear interpretation: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [8] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [0] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] We have applied the subprocessor on the resulting DP-problem: 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {p^#(s(x1)) -> c_5()} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} Details: We apply the weight gap principle, strictly orienting the rules {p^#(s(x1)) -> c_5()} and weakly orienting the rules { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {p^#(s(x1)) -> c_5()} Details: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [1] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'Empty TRS' ----------- Answer: YES(?,O(1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {} Weak Rules: { p^#(s(x1)) -> c_5() , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} Details: The given problem does not contain any strict rules 10) {j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} The usable rules for this path are the following: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} We have oriented the usable rules with the following strongly linear interpretation: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [8] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [0] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] We have applied the subprocessor on the resulting DP-problem: 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} Details: We apply the weight gap principle, strictly orienting the rules {j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} and weakly orienting the rules { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1))))))))))))} Details: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [1] x1 + [1] c_2(x1) = [1] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'Empty TRS' ----------- Answer: YES(?,O(1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {} Weak Rules: { j^#(0(x1)) -> c_2(p^#(s(p(p(s(s(0(p(s(p(s(x1)))))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} Details: The given problem does not contain any strict rules 11) { i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1)))))))))) , p^#(s(x1)) -> c_5()} The usable rules for this path are the following: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} We have oriented the usable rules with the following strongly linear interpretation: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [8] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [0] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] We have applied the subprocessor on the resulting DP-problem: 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {p^#(s(x1)) -> c_5()} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} Details: We apply the weight gap principle, strictly orienting the rules {p^#(s(x1)) -> c_5()} and weakly orienting the rules { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {p^#(s(x1)) -> c_5()} Details: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [1] x1 + [1] c_0(x1) = [1] x1 + [0] p^#(x1) = [1] x1 + [1] c_1(x1) = [0] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'Empty TRS' ----------- Answer: YES(?,O(1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {} Weak Rules: { p^#(s(x1)) -> c_5() , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1))))))))) , i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} Details: The given problem does not contain any strict rules 12) {i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} The usable rules for this path are the following: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} We have oriented the usable rules with the following strongly linear interpretation: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [8] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [0] x1 + [0] c_0(x1) = [0] x1 + [0] p^#(x1) = [0] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] We have applied the subprocessor on the resulting DP-problem: 'Weight Gap Principle' ---------------------- Answer: YES(?,O(n^1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} Weak Rules: { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} Details: We apply the weight gap principle, strictly orienting the rules {i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} and weakly orienting the rules { p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} using the following strongly linear interpretation: Processor 'Matrix Interpretation' oriented the following rules strictly: {i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1))))))))))} Details: Interpretation Functions: i(x1) = [0] x1 + [0] 0(x1) = [1] x1 + [0] p(x1) = [1] x1 + [0] s(x1) = [1] x1 + [0] j(x1) = [0] x1 + [0] i^#(x1) = [1] x1 + [1] c_0(x1) = [1] x1 + [0] p^#(x1) = [1] x1 + [0] c_1(x1) = [0] x1 + [0] j^#(x1) = [0] x1 + [0] c_2(x1) = [0] x1 + [0] c_3(x1) = [0] x1 + [0] c_4(x1) = [0] x1 + [0] c_5() = [0] c_6() = [0] Finally we apply the subprocessor 'Empty TRS' ----------- Answer: YES(?,O(1)) Input Problem: innermost DP runtime-complexity with respect to Strict Rules: {} Weak Rules: { i^#(0(x1)) -> c_0(p^#(s(p(s(0(p(s(p(s(x1)))))))))) , p(p(s(x1))) -> p(x1) , p(s(x1)) -> x1 , p(0(x1)) -> 0(s(s(s(s(s(s(s(s(x1)))))))))} Details: The given problem does not contain any strict rules