Research And Implementation Of Online Game Agent Based On BDI Model

In recent years,with the rapid social and economic development,online game market develops with an increasing rate due to the rising living standard and spare time.As a kind of supplementary element of online game,agent plays an important role in improving game interest and intelligence,has attracted more and more attentions.Homogeneity is a common problem in the process of online game agent design,and leads to agent's lack of personality characteristics.This paper aims at this problem.In particular,From the theoretical study perspective,this paper investigates the reasoning process of BDI agent under online game background:(1)For the classic BDI agent reasoning model,extend the Belief set with personality elements,endowing agent with individualized feature;(2)In the evaluation of agent target,the subjective factors such as personality elements and emotions are combined with objective factors such as value and difficulty,and the expectation degree is used as the target evaluation result.The personality element is used to determine the basic point of the proportion of the value and difficulty in the target evaluation,and the deviation of the evaluation is determined by the emotion,so that the target evaluation process is closer to man decision;(3)In the design of emotional behavior mechanism,the transformation relation model between event,personality,emotion,target and behavior is established,which provides the basis for judging the behavior of online game agent.From the practice perspective,this paper applied the proposed agent model to the design of one kind of Mahjong game agent:(1)Instantiate components of BDI game agent namely Belief,Desire and Intention;(2)Realize the BDI based online game agent using the existing online game platform,and analyze the influence of personality elements and emotions on the behavior of online game agent.This paper makes a theoretical discussion on the design of personalized online game Agent,and provides a theoretical idea for the method of behavior generation in card game agent design.