A computational-intelligence-based optimization of agents' initial positions in land combat simulation *

Abstract

Abstract The optimization of agents' initial properties enables agents to perform their assigned tasks more perfectly. This paper presents an optimizing method using the combination of radial basis function (RBF) neural network and genetic algorithm (GA). In the land combat simulation, the method can ensure that the agents optimized defeat the agents not optimized absolutely. Compared with the optimization based on support vector machines (SVM), the proposed method improves the efficiency more than twenty times, so it suits the cases where the speed as well as performance is required.