Adaptive Increasing/Decreasing PSO for Solving Dynamic Optimization Problems

Science progress has introduced new issues into the world requiring optimization algorithm with fast adaptation with uncertain environment changing with time. In these issues, location and optimized value change over time, so optimization algorithm should be capable of fast adaptation with variable conditions. This study has proposed a new algorithm based on particle optimization algorithm called Adaptive Increasing/Decreasing PSO. This algorithm, adaptively with an increase and decrease in the number of algorithm particles and effective search limit, is capable of searching and finding optimized number changed with time in non-linear and dynamic environments with undetectable changes. Also, a new definition, focused search zone, is provided for signalizing hopeful areas in order to accelerate local search process and prevent premature convergence, and success index as an indicator of the behavior of centralized search area in relation to environmental conditions. Results of the proposed algorithm on the moving peaks benchmark were assessed and compared with the results of some other studies. Results show positive effects of adaptive mechanisms such as a decrease and an increase in the particles and search limit on the duration of searching and finding optimization in comparison with other multi-population based optimization algorithms.