An Improved Fruit Fly Optimization Algorithm Incorporating Average Learning and Step Changing

Abstract

To overcome the disadvantages of slow convergence speed, easily relapsing into local extremum and poor stability of traditional fruit fly optimization algorithm (FOA), an improved FOA incorporating Average Learning and Step Changing into the evolutionary process (AL-SC-FOA) is proposed. The combination of strategies of average learning and step changing can balance the ability of global search and local search of the algorithm, accelerate the convergence speed, improve the accuracy and enhance the stability of the algorithm. The proposed AL-SC-FOA is applied to six standard examples of the optimization test. The experimental results show that the AL-SC-FOA can avoid falling into the local optimum, which has higher precision and faster convergence speed, as well as better stability.