A novel artificial bee colony algorithm with local and global information interaction

The artificial bee colony algorithm (ABC) is a new stochastic and population-based optimization method, which has been attracting a great deal of attention, due to its simple structure, easy implementation and outstanding performance. However, it also suffers from slow convergence like other evolutionary algorithms. In order to address this concerning issue, in this paper, we propose a novel artificial bee colony algorithm with local and global information interaction, called ABCLGII. In employed bee phase, each employed bee is designed to learn from the best individual among its neighbors or in a local visible scope. By this way, the search of employed bees is no longer independent and blind, but is cooperative and directional, such that a local information interaction mechanism is conducted between employed bees. In onlooker bee phase, only a part of superior food sources have chance to attract onlooker bees to exploit in their vicinity. Moreover, two novel search equations are proposed for onlooker bees to generate candidate food sources. Specifically, one exploits the useful information of some good solutions, while the other combines the valuable information of the current best solution and some good solutions simultaneously. An adaptive selection mechanism is accordingly designed for onlooker bees to choose a proper search equation for producing candidate food sources. In this way, a global information interaction mechanism is employed for onlooker bees. In order to evaluate the performance of ABCLGII, we compare ABCLGII with the original ABC and other outstanding ABC variants on 52 frequently used test functions. The experimental results show that ABCLGII is better than or at least competitive to the state-of-the-art ABC variants in terms of solution quality, robustness and convergence speed.