A modified artificial bee colony algorithm for load balancing in network-coding-based multicast

Abstract

This paper studies the load balancing optimization problem in network-coding-based multicast and proposes a modified artificial bee colony algorithm (MABC) to address it. MABC is featured with three novel schemes, including a food source initialization scheme, a novel selection scheme and a neighborhood search scheme. The first scheme generates a set of high-quality food source positions, ensuring that the exploration of the search begins with promising areas in the search space. In the second scheme, a nectar source library (NSL) is used to store a set of best solutions found during the iterative search. Each scout bee produces a new food source based on a food source randomly selected from NSL. This helps to generate food sources with high nectar amounts. The last scheme is a neighborhood search scheme to strengthen population diversity and avoid local optima, where a probability vector is maintained and utilized to carry out fine local exploitation. Experimental results demonstrate that the proposed MABC outperforms a number of state-of-the-art evolutionary algorithms with respect to the quality of solutions obtained.