The Simulation Of Group Behavior Based On Cellular Automata

This paper establishes the model of group behavior which based on cellular automata, anduses cellular automata simulate group behavior. We describe group behavior according to birdsparticles, employees and area businesses this three levels of specific examples. After simulating,we eventually attempt to reveal the internal operating mechanism of group behavior.Firstly, this paper establishes the model of particle swarm optimization based on cellularautomata, and then uses it solve benchmark functions. We observed that PSO algorithm was easilyfalling into local extreme point by using cellular automata simulation, and developed severalimprovement strategies which drew inspiration from simulation, such as adding additional particleinto updated formula, introducing crossover operator and mutation disturbance when the prematureconvergence occurs. Our experiments validated new algorithm on benchmark functions, andresults showed that the improved PSO algorithm has a strong group search capability.Secondly, employees are simplified to cellular automata with different properties. This paperestablishes the model of employee behavior based on cellular automata. State, moving rules andtransfer algorithm of cells are also proposed in the model. The results of simulation show thatloyalty and cohesion for a single and hybrid staff increases at first and stabilizes after a period oftime. Simulation for the controllable incentive management illustrates that positive incentive isimplemented at first improves employee loyalty and cohesion among economic staffs. Through thenumber of iteration increasing, loyalty and cohesion from the two groups are equal final. On theother hand, among the hybrid staffs, the cohesion of group with negative incentive implemented atfirst is equal with the group with positive incentive implemented all the time.Thirdly, This paper introduces three essential characters of the enterprise’s behavior intocellular automata model, which contains the demand character of the enterprise itself, the changefactor of surrounding neighborhood and the memory attribute of the enterprise. In addition,according to the evolution rules and the neighborhood structure, the individual business moves onestep or waits in the original position. The results of simulation show that in the case of taking anaverage of four preference parameters, aggregation of enterprise increases with a rising ofindividuals number. In addition, experiments also show that all three essential characters have acertain influence on individual behavior as well as group’s.