ESTIMATION DISTRIBUTION ALGORITHM ASSOCIATED WITH MEMORY SCHEMES IN DYNAMIC SEARCH SPACES

Abstract

The Evolutionary Algorithms (EA) are target of major studies regarding their use in dynamic optimization problems. This observation reflects the real importance of AE in real world applications. Many studies have been developed in order to have a better understand over the memory schemes. This work investigates the application of Multi-PBIL method on dynamic search spaces, using memory schemes. The Multi-PBIL is an optimization algorithm that belongs to the class of evolutionary algorithms. An adaptation is proposed to improve its adaptability in dynamic environments. In memory schemes, the probability models are stored in memory, along with the best samples created in a search space and they are used to reactivate old environments from the point when changes occur. Experimental studies based on a series of dynamic environments shows the efficiency of the memory scheme for the proposed method. This work presents also the relationship between the memory scheme and the multi-population in dynamic environments. Keywords: Estimation Distribution Algorithm, Dynamic Optimization Problem, Scheme Memory, Multi-PBIL, Evolutionary computation.