One-dimensional blind-walking optimization method and application

In order to improve the optimizing efficiency, one-dimensional blind-walking optimization method is proposed, which can be realized by halving step or doubling step for the applicability test condition of the sensing point. The optimum point, then, can be reached at high rate by doubling step and converged by halving step. Current point should be updated in whole design space. The flow chart of this algorithm with operating process is put forward. And then, two optimization problems with unimodal and multimodal objective functions are solved respectively. The simulation results show that the proposed method is better than the ordinary method, which has the advantages of fast convergence speed, less calculating amounts, and wide application scope, etc. Taking the method as innovative kernel, random research method, feasible direction method and other complex methods are improved. Its characteristics are suitable for vivid teaching like interpreting the penalty function method as wall mound method.