Multi-objective Route Planning Based on Improved K-means Algorithm

Abstract

Multi-objective route planning is a hot issue in current research, and it applies all aspects of life. With the expansion of the scale of the problem, large numbers of approximate algorithms and heuristic algorithms proposed to solve the problem. In this paper, a solution of a multi objective route planning with a balanced assignment of tasks is proposed. The solution can divide into two steps. First., a clustering algorithm cbk-means (cluster balance k-means) is proposed, which improves the similarity measurement in the clustering process, and overcomes the shortcomings of traditional k-means algorithm, such as uncertain number of points and inflexible measurement criteria, which is the key step to achieve fair assignment of tasks. Second, this paper use genetic algorithm to obtain an optimal route planning for each cluster. Experimental results show that the cbk-means algorithm makes the workload of each cluster more balanced at the expense of negligible cost, which improves the fairness of task assignment greatly. Besides, this hybrid solution can save computational time and get better results.