Solving Network Flow Planning and Service Network Design problems due to their large set of variables is difficult. Fig. 3 presents the solution methodologies used, and clearly shows that due to complexity of these problems, heuristic and metaheuristic solution methods are the prime choice. Among them, Ta-bu Search seems to be a popular meta-heuristic algorithm. However, we found no paper applying approximation techniques, which is an interesting research opportunity.
In order to evaluate the performance of the propose algorithms, good benchmark sets are needed. The majority of the research on tactical planning level address specific real-world problems and the algorithms designed for those problems. Thus, a solid analysis on the efficiency of different algorithms, due to the lack of general benchmark sets, is not attainable. We can only mention some of the interesting results obtained by the researchers. For instance, Bektas, Chouman, and Crainic (2010) compare arc-based and flow-based decomposition methods in solving small sizes of their non-linear problem, and show that despite the fact time compared to flow-based decomposition. In overall, both techniques handle the non-linearity efficiently