Title: Adaptive Memory Matrices for Automatic Termination of Evolutionary Algorithms

Year of Publication: Jul - 2015
Page Numbers: 1-11
Authors: Abdel-Rahman Hedar
Conference Name: The Fourth International Conference on Informatics & Applications (ICIA2015)
- Japan


Evolutionary Algorithms (EAs) still have no automatic termination criterion. In this paper, we modify a genetic algorithm (GA), as an example of EAs, with new automatic termination criteria and acceleration elements. The proposed method is called the GA with Gene and Landmark Matrices (GAGLM). In the GAGLM method, the Gene Matrix (GM) and Landmark Matrix (LM) are constructed to equip the search process with a self-check to judge how much exploration has been done and to maintain the population diversity. Moreover, a special mutation operation called “Mutagenesis” is defined to achieve more efficient and faster exploration and exploitation processes. The computational experiments show the efficiency of the GAGLM method, especially its new elements of the mutagenesis operation and the proposed termination criteria.