Abstract:

This paper presents an improved GSO algorithm. The modification has three main aspects: firstly, uses an improved sharing strategy which shares not only the best member but also the other good members’ information; secondly, uses completely random search instead of search by the head angle; thirdly, restricts the number of changing dimensions of the scroungers. Among them, the first one is proposed in this paper. When tested against six benchmark functions, in low and high dimensions, the improved algorithm has a  competitive performance to GSO, GA and PSO, especially on highdimensional problems.

Key words: optimizer algorithm;swarm intelligence;GSO algorithm;function optimization