Model for Human Killer Solving

Abstract

I research how Killer Sudokus are solved by humans, primarily through introspection. I settle on modelling the process with a Priority Queue for steps, or Rules. I thoroughly examine these Rules and Killer Sudokus in general. The model has merit, for it is able to solve most puzzles of low to medium difficulty and can represent varying preferences for the order of the solving process. It even allows for emulating logical errors humans might make by changing priorities. The model is lacking in that its ruleset is not yet strong enough to solve problems that experienced humans can solve. The approach seems sound, but requires refinement and additional work.