基于多目标进化的终端聚合选择算法研究

摘要

When the service capacity of a terminal is limited, how to aggregate heterogeneous terminals according to their capacity to support service application is a challenge. This paper proposes a terminal aggregative selection algorithm based on Multi-Objective Evolutionary (MOE), describes the framework of terminal aggregation and designs 0-1 programming model. The proposed algorithm determines automatically the weights of multi-index, avoiding the difficulty to specify the weights manually. Simulation results show that the proposed algorithm can produce a set of optimal Pareto solutions and the most approving solution is selected as the final decision according to preference and application requirements. Also the proposed algorithm provides an improvement to the feasibility and efficiency.%当单一终端业务能力受限时,如何根据终端能力对异构终端进行有效聚合,以支持内容丰富的业务应用成为一个非常重要的问题.该文提出一种基于多目标进化的终端聚合选择算法,给出终端聚合的描述框架,构建基于个体和协同能力的0-1规划模型,包括多指标权重自动生成法,减轻用户负担,避免用户直接指定权重的困难.仿真结果表明,算法能得到一个Pareto最优解集,用户根据其对个体和协同表现的个人偏好或实际业务应用需要选择最满意的方案聚合终端,同时算法具有良好的可行性和有效性.