Application of PSO and GA for Transmission Network Expansion Planning

TNEP is one of the important parts of power system planning which determines the number, time, and location of new lines for adding to transmission network. It is a hard, large-scale and highly nonlinear combinatorial optimization problem that can be solved by classic, nonclassic or heuristic methods. Classic methods like linear programming and Bender decomposition are only based on mathematical principles, but their difficulty is that if the scale of problem is large, it is very difficult to find accurate and reasonable solutions. Contrary to classic methods, nonclassic ones such as evolutionary algorithms like GAs are not based on mathematical rules and simply can be applied for solution of complex problems. GA is a random search method that has demonstrated the ability to deal with nonconvex, nonlinear, integer-mixed optimization problems like the STNEP problem. Although global optimization techniques like GA to be good methods for the solution of TNEP problem, however, when the system has a highly epistatic objective function and number of parameters to be optimized is large, then they have degraded efficiency to obtain global optimum solution and also simulation process use a lot of computing time. Heuristic methods like PSO can improve speed and accuracy of the solution program. PSO is a novel population-based heuristic that is a useful tool for engineering optimization. Unlike the other heuristic techniques, it has a flexible and well-balanced mechanism to enhance the global and local exploration abilities. In this chapter, first we review some research in the field of TNEP. Then, the method of mathematical modeling for TNEP problem is presented. Afterward, GA and PSO algorithms are described completely. Finally, effective parameters on network losses with a few examples are introduced.