论文信息 - A parallel GA-BP neural network algorithm for solving inverse heat conduction problem

A parallel GA-BP neural network algorithm for solving inverse heat conduction problem

Inverse heat conduction problem (IHCP) is a nonlinear, ill-posed problem of huge calculation, which is difficult to be solved properly. This paper proposes a parallel GA-BP algorithm on cluster for solving IHCP, which combine the ability of nonlinear mappings of forward back propagation (BP) neural network and the capacity of global optimization of genetic algorithm (GA). The processes of optimization algorithm and data parallelism are introduced mainly in this paper. The experimental results show that the method can achieve high speedup and improve computational efficiency.

Yan Chen | Qingping Guo | Li Ma

[1] Y. Hwang,et al. Applying neural networks to the solution of forward and inverse heat conduction problems , 2006 .

[2] Shesheng Zhang,et al. Optimum tactics of parallel multi-grid algorithm with virtual boundary forecast method running on a local network with the PVM platform , 2008, Journal of Computer Science and Technology.

[3] Ron Shonkwiler,et al. Parallel Genetic Algorithms , 1993, ICGA.

[4] Shesheng Zhang,et al. Parallel Multi-grid Algorithm with Virtual Boundary Forecast Domain Decomposition Method for Solving Non-linear Heat Transfer Equation , 2000, HPCN Europe.

[5] Yu Shengsheng. Algorithm of BP Neural Network-based Face Detection , 2004 .

[6] Xin Liu,et al. A dynamic genetic algorithm based on continuous neural networks for a kind of non-convex optimization problems , 2004, Appl. Math. Comput..