The problematic nature of potentially polynomial-time algorithms solving the subset-sum problem
share
The main purpose of this paper is to study the NP-complete subset-sum problem, not in the usual context of time-complexity-based classification of the algorithms (exponential/polynomial), but through a new kind of algorithmic classification which we introduce based on a property that all known exponential-time algorithms share. We then construct a theoretical mathematical environment within which we compare the two classes that are produced from the new classification; one class is characterized by a normal mathematical nature, whereas the other by a problematic one. These results are transferred to exponential/polynomial algorithms, through a conjecture that links the two classifications. As for the mathematical environment, it consists of a simple random experiment designed in such a way that the algorithmic operation is linked to it. We study the random experiment with a stochastic process which is the main tool for the comparison of the two classes.
READ FULL TEXT
