为什么类 P 和 NP 在有限性意义上并非良定义
Why the classes P and NP are not well-defined finitarily
AI总结:
本文通过区分算法可验证性与算法可计算性,论证哥德尔式命题可验证但不可计算,并据此以算术方式解释 PvNP 问题,得出 P 和 NP 并非有限性良定义且 SAT 既不属于 P 也不属于 NP 的结论。
摘要:
我们在有限性意义上区分算法可验证性与算法可计算性,以表明哥德尔的“形式上”不可证明、但“按数值逐项”可证明的算术命题 [(Ax)R(x)] 可以被有限性地证明为:作为“始终”为真在算法上可验证,但作为“始终”为真在算法上不可计算。因此,尽管 [R(x)] 作为重言式在算法上可验证,但它不能被任何图灵机(无论是确定性的还是非确定性的)作为重言式在算法上计算。通过以算术方式而非集合论方式解释 PvNP 问题,我们得出结论:类 P 和 NP 在有限性意义上并非良定义,因为由此立即得出 SAT 既不在 P 中也不在 NP 中。
英文摘要:
We distinguish finitarily between algorithmic verifiability, and algorithmic computability, to show that Goedel's 'formally' unprovable, but 'numeral-wise' provable, arithmetical proposition [(Ax)R(x)] can be finitarily evidenced as: algorithmically verifiable as 'always' true, but not algorithmically computable as 'always' true. Hence, though [R(x)] is algorithmically verifiable as a tautology, it is not algorithmically computable as a tautology by any Turing machine, whether deterministic or non-deterministic. By interpreting the PvNP problem arithmetically, rather than set-theoretically, we conclude that the clkasses P and NP are not well-defined finitarily since it immediately follows that SAT is neither in P nor in NP.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。