摘要
This paper is concerned with two important issues in multicast routing problem with network coding for the first time, namely the load balancing and the transmission delay. A bi-objective optimization problem is formulated, where the average bandwidth utilization ratio and the average transmission delay are both to be minimized. To address the problem, we propose a novel multiobjective artificial bee colony algorithm, with two performance enhancing schemes integrated. The first scheme is an elitism-based food source generation scheme for scout bees, where for each scout bee, a new food source is generated by either recombining two elite solutions randomly selected from an archive or sampling the probabilistic distribution model built from all elite solutions in this archive. This scheme provides scouts with high-quality and diversified food sources and thus helps to strengthen the global exploration. The second one is a Pareto local search operator with the concept of path relinking integrated. This scheme is incorporated into the onlooker bee phase for exploring neighboring areas of promising food sources and hence enhances the local exploitation. Experimental results show that the proposed algorithm performs better than a number of state-of-the-art multiobjective evolutionary algorithms in terms of the approximated Pareto-optimal front.
This paper is concerned with two important issues in multicast routing problem with network coding for the first time, namely the load balancing and the transmission delay. A bi-objective optimization problem is formulated, where the average bandwidth utilization ratio and the average transmission delay are both to be minimized. To address the problem, we propose a novel multiobjective artificial bee colony algorithm, with two performance enhancing schemes integrated. The first scheme is an elitism-based food source generation scheme for scout bees, where for each scout bee, a new food source is generated by either recombining two elite solutions randomly selected from an archive or sampling the probabilistic distribution model built from all elite solutions in this archive. This scheme provides scouts with high-quality and diversified food sources and thus helps to strengthen the global exploration. The second one is a Pareto local search operator with the concept of path relinking integrated. This scheme is incorporated into the onlooker bee phase for exploring neighboring areas of promising food sources and hence enhances the local exploitation. Experimental results show that the proposed algorithm performs better than a number of state-of-the-art multiobjective evolutionary algorithms in terms of the approximated Pareto-optimal front.
基金
supported in part by National Natural Science Foundation of China (No.61505168, No. 61802319)
the Fundamental Research Funds for the Central Universities, P. R. China