Comparative strategies for knowledge migration in Multi Objective Optimization Problems
Panth Parikh, Ziad Kobti · 2017
This paper provides a comparative study of the knowledge migration strategies in multi-objective optimization problems. It offers different migration strategies which are inspired by the game theory model. The idea behind incorporating this strategy is to increase diversity among the population, avoid premature convergence and escape from local optima. It also provides a meaningful migration in context to the population environment. Migration can be according to the individual choice, the decision of best individuals in the subpopulation or by negotiation among the population. The strategies used are of economics background which include the social factor which makes the individuals use their knowledge and decide the region of migration. This allows the population to explore new regions in the search space and increase diversity as the migrating individual carries its culture knowledge to other populations. The proposed algorithm is tested against CEC 2015 expensive benchmark problems. Results depict that it leads to better performance when migration is carried out by making the use of the proposed strategy.