A Simplex-based simulated annealing algorithm for node-arc capacitated multicommodity network design

The network design is a well-known problem, both of practical and theoretical significance. Network design models are extensively used to represent a wide range of planning and operations management issues in transportation, telecommunications, logistics, production and distribution. This paper presents a solution method for node-arc formulation of capacitated fixed-charge multicommodity network design problems. The proposed method is a hybrid algorithm of Simplex method and simulated annealing metaheuristic. The basic idea of the proposed algorithm is to use a simulated annealing algorithm to explore the solution space, where the revised Simplex method is used to evaluate, select and implement the moves. In the proposed algorithm, the neighborhood structure is pivoting rules of the Simplex method that provide an efficient way to reach the neighbors of current solution. To evaluate the proposed algorithm, the standard problems with different sizes are used. The algorithm parameters are tuned by design of experiments approach and the most appropriate values for the parameters are adjusted. The performance of the proposed algorithm is evaluated by statistical analysis. The results show high efficiency and effectiveness of the proposed algorithm.

Figure optionsDownload as PowerPoint slideHighlights
► Proposing an efficient algorithm to solve the capacitated network design problem.
► Presenting a DOE-based method to tune the parameters of the algorithm.
► Providing experimental results on a large set of standard test problems.
► Proving superiority of the algorithm by using statistical analysis.