Abstract:

Dynamic optimization problems are widespread in actual production or life, and environmental detection and response methods are the core of solving such problems. In many practical problems, due to the interference of random factors, the true optimal solution of the optimization problem will be randomly offset to a certain extent. This paper considers the stochastic dynamic optimization problem in which the random offset of the optimal solution obeys the normal distribution. First of all, this paper improves the existing interval shrinkage method based on the idea of orthogonal experimental design, and then proposes an environmental detection and response strategy for dynamic optimization problems, which avoids the blindness and randomness of the existing methods to a certain extent. Secondly, the upper limit of the standard deviation of the random perturbation corresponding to no change in the environmental detection before and after the perturbation is given. Finally, the particle swarm optimization algorithm is used for testing. And the experimental results show that the environmental detection and response method proposed in this paper can not only effectively deal with the stochastic dynamic optimization problem in which the optimal solution is disturbed by random, but also improve the ability of using particle swarm optimization to deal with other dynamic optimization problems. The improved environment detection and response method can be applied to other evolutionary algorithms besides particle swarm optimization.

Key words: Dynamic optimization, Environmental detection and response, Random disturbance, Orthogonal experimental design