Aplicação de planejamento em jogos de estratégia

Abstract

With a digital game industry in constant ascendant and a more demanding consumer, designers and programmers have faced even bigger challenges to create realistic games. In this way, people are working a lot in the NPCs (Non Player Characters) intelligence. In this research, it was emphasized the use of planning to control the Artificial Intelligence of the NPCs. The work documented here was guided by the verification of two hypotheses: the viability of using a planner based on the A* algorithm and on a knowledge representation model similar to STRIPS system in strategic games, and the possibility of using this planner in a Multiagent System approach based on an auction system used to identify agents more appropriated to fulfill a goal. The term strategic games was chosen to avoid misunderstandings about the traditional RTS games. The planner and the Multiagent System have been implemented and several tests were made. The results were satisfactory and the hypotheses proved.