A modified fruit fly optimization algorithm with better balance between exploration and exploitation

Abstract

This paper presents a modified fruit fly optimization algorithm(FOA). The proposed modified FOA establishes a balanced tradeoff between exploration and exploitation, and thus overcomes original FOA's drawbacks of premature convergence and easy trapping in a local optima. In the proposed modified FOA, firstly, the whole population performs a global search; Secondly, the whole population are sequenced in descending order by the individual fitness value; Thirdly, every n consecutive individuals are divided into a meme group, then every meme group iteratively performs a deep search around the local optima; Finally, all the meme groups are mixed, and then the above process is implemented iteratively until meeting the end conditions. The modified FOA efficiently avoids relapsing into local optima and improves convergence precision. Finally, our modified algorithm was validated against the original by testing on six standard benchmark functions, and comparisons show that the performance of the proposed modified FOA is much better than original FOA.