Performance improvement in game playing using evolutionary computation by large search space exploration

Soft Computing branch of intelligence research is primarily focused on the path of achieving high performance by mimicking the human approach. The central idea is to capture and encode human knowledge in artificial learning form. Applying AI technology to develop efficient game-playing programs is through realization of search-intensive approach in very large and complex search intensive areas. Many researchers have developed very powerful search techniques over the past two decades and successfully applied these search algorithms to problems domains of optimization, machine learning and soft computing paradigms. This paper extends this approach, by developing a program which is almost completely reliant on search optimization through evolutionary computation. Very efficient evolutionary algorithms and advancement of “intelligent” search along with improved hardware resources like faster processors, larger memories, and larger disks makes it possible to push the limits to solve problem of type and size of Checkers which has search space as high as 5 × 10²⁰ representing a daunting challenge. The collected checkers result pushes the boundary of evolutionary algorithms based problem domains.