The Curry-Howard correspondence is often called the proofs-as-programs result. I offer a generalization of this result, something which may be called machines as programs. Utilizing this insight, I introduce two new Turing Machines called "Ceiling Machines." The formal ingredients of these two machines are nearly identical. But there are crucial differences, splitting the two into a "Higher Ceiling Machine" and a "Lower Ceiling Machine." A potential graph of state transitions of the Higher Ceiling Machine is then offered. This graph is termed the "canonically nondeterministic solution" or CNDS, whose accompanying problem is its own replication, i.e., the problem, "Replicate CNDS" (whose accompanying algorithm is cast in Martin-L\"of type theory). I then show that while this graph can be replicated (solved) in polynomial time by a nondeterministic machine -- of which the Higher Ceiling Machine is a canonical example -- it cannot be solved in polynomial time by a deterministic machine, of which the Lower Ceiling Machine is also canonical. It is consequently proven that P $\neq$ NP.
翻译:Curry- Howard 函文通常被称为“ 证明- 程序” 结果。 我提供这种结果的概略化, 可能是机器作为程序。 利用这个洞察力, 我引入了两个新的图灵机器, 称为“ 超高机器 ” 。 两台机器的正式成分几乎完全相同。 但是, 存在一些关键差异, 将两者分成“ 更高的天花板” 和“ 更低的天花板 ” 。 然后提供了高天花机( 高天花机) 或 CNDS 的潜在国家过渡图。 这个图称为“ 具有非决定性的非决定性解决方案 ” 或 CNDS, 其附带的问题是它自身的复制, 即问题, “ 复制 CNDS ” ( 其伴随的算法在类型理论中的 Martin- L\\\ ” ( ) 。 然后我显示, 虽然这个图可以通过非定时的机器复制( 溶性机器) 。 高天花机是一个典型的例子。 这个图无法在多明时间里解决它, 。 。