On path selection for multipath connection
Department of Applied Computing
Multipath connection, which utilizes the multiple paths between network hosts in parallel, has been used to improve the network performance, security and reliability. Path selection is a critical decision in a multipath connection network. Different selection results in significantly different result. In this paper, we present several heuristic algorithms including genetic algorithm to solve the path selection problem in a multipath connection environment. The reasons to choose genetic algorithm are because of its flexibility and extensibility when the context of problem changes. We define two objective functions and two constrains in this problem. The performance results of the proposed algorithms on the simulated network topology as well as a real-world network topology are presented. It is observed that genetic algorithm can produce satisfactory results within reasonable execution time.
Advances and Innovations in Systems, Computing Sciences and Software Engineering
Edward Chow, C.
On path selection for multipath connection.
Advances and Innovations in Systems, Computing Sciences and Software Engineering, 513-518.
Retrieved from: https://digitalcommons.mtu.edu/michigantech-p/4026