Optimal placement and deployment strategies in mobile agent-based network management

Abstract

The paper aims to investigate how to improve the performance of a mobile agent-based network management system. We achieve this goal through finding optimal placement and deployment strategies. First, we formulate the problem of finding the agent server's optimal placement strategy as a vertex cover (VC) problem that is known to be NP-hard, and provide a hybrid genetic algorithm to solve it. This algorithm introduces a new indirect encoding based on edges and an extended mutation operator. Comparing with other algorithms, we find that our algorithms are superior in the aspects of performance and efficiency. Second, we abstract the searching of an agent's optimal deployment strategy to a variation of the vehicle routing problem that is also NP-hard, and propose a two level genetic algorithm to obtain optimal solutions.