부산대학교 도서관

Many computer applications require solving difficult problems (NP or PSPACE), which can be expensive to analyze the space of feasible solutions. In the last 20 years, the theoretical computer science area showed enormous progress in the algorithm design that is not being put in practice. Experimental algorithmic can help to bridge the gap between theory and practice. A domain of applications that involves hard problems is the development of computer games, in particular, games in which the user's opponent is the computer. Heuristic search technique in games has been highly successful, however, there are no tools to assist in testing and improving implementations based on it. In this paper, we describe a framework called GLIESE that focuses on the reuse of software components for experimental game development. Its main goal is to provide software support to simplify the most complex issues in the implementation of video games enabling to experiment with games in which parts of its logic can be expressed by using different heuristic searches. The simulation and assess of the quality of approximate solution based on different heuristic search techniques is the original aspect that impacts on the game implementation. GLIESE was used in undergraduate courses of "Analysis and Design of Algorithms" in the development of variety of popular games.