Based On Decision Theory, Particle Swarm Optimization

Particle swam optimization (PSO) is an intelligence optimization algorithm by simulating the animal swarm behavior. From other point of view, it also simulates the human society behavior; however, there are seldom research reports for this aspect. Therefore, this paper introduces the individual decision-making theory in which each particle is viewed as one agent and has independent individual decision-making ability, and applied it into solving nonlinear equations.PSO uses past experiences and other members experiences to adjust their behavior, but it has some inadequate in using individual historical experience information. Therefore,with the help of individual decision-making theory and method,we make decision to the individual historical optimization position of stand PSO using individual historical position and corresponding fitness.Due to human society is a complicated society system,PSO is introduced small-word modal and individual decision-making theory and way,each particle has special topological neighborhood information,these change the abuse of stand PSO only use the swam history optimization.Threfore we make decision individual history optimization position using the particles neighborhood individual historical position and corresponding fitness.Finally, using Lyapunov function theoretical analyze the stability conditions of two improved SPO.Cognitive coefficient and social coefficient are always set a fixed number or set the same at the same genetion.As agent, each particle has the ability of individual decision-making, and therefore the two parameters should have some difference. In this paper, a new judged stand is designed for particles evaluate (good or bad)—fitness change of rate. We make decision cognitive coefficient and social coefficient using the particles individual history position and corresponding fitness with the help of individual decision-making and fitness change of rate.At last, individual decision-making PSO is applied into solving nonlinear equations problem, simulation results show they are more superior.