Application Of GPU Parallel Algorithm In The Solution Of Problems In The Game Theory Of Economics

In order to improve the intelligent swarm algorithm in the body of economics game numerical experiments in computational efficiency and to solve the game number or number of strategies effectively,under the condition of calculation of the proposed particle swarm PSO algorithm based on approximate Nash equilibrium,and the combination of GPU technology and CUDA algorithm calculation tool,to solve algorithm of complete information static game strategy under the parallel implementation of reconstruction.The reconstruction mainly depends on the Dynamic-parallel,Unified-address,Stream-flow,CUDA parallel computing technology,using data flow decomposition algorithm and block matrix operations such as parallel computing skills;and puts forward the reconstruction algorithm and continuous discretization strategy based on the combination of approximation algorithm,solved the Nash equilibrium exists in a continuous game only pure strategy equilibrium approximation problem.Through numerical simulation,the Cournot game sewage economics game model,demonstrated with the help of the GPU by increasing the number of particles can effectively alleviate the PSO algorithm can not only guarantee the convergence to the global optimum,and the local optimum problem,but demonstrates the feasibility of parallel PSO algorithm reconstructed with better strategies and efficiency in solving the game or having more game model.