Effect of negative campaign strategy of election algorithm in solving optimization problem

Abstract

Election algorithm (EA) is an optimization technique based on minimization and coalition operations to solve competition among neurons. The Election algorithm gives the best individual of the population by enhancing both minimization and coalition operations while local search gives the best local solutions by testing all neighbouring solutions. Negative campaign mechanism is one of the most important mechanism in EA for its impact on the diversification and overcoming premature convergence of the entire search space towards optimal searching. The challenging task lies in selecting the appropriate negative campaigning operator that leads to optimal searching in a reasonable amount of time. The decision then becomes more difficult and needs more trial and error to find the best negative campaigning operator. This paper investigates the effect of negative campaign operators in enhancing the performance of EA based on the Travelling Salesman Problem (TSP). New negative campaign operator has been proposed based on selecting the best voter to be replaced. Experiments were conducted on the TSP to evaluate the proposed methods. The proposed mechanism was compared with other negative campaign operators. The result reveals the significant enhancement of the EA performance based on the proposed method in TSP problem.