Abstract—This paper presents the application of an efficient shuffled frog leaping algorithm (SFLA) to solve the optimal generation expansion planning (GEP) problem. The SFLA is a meta-heuristic search method inspired by natural memetics. It combines the advantages of both genetic-based memetic algorithms and social behavior based algorithm of particle swarm optimization. Least-cost GEP is concerned with a highly constrained non-linear discrete dynamic optimization problem. In this paper the proposed formulation of problem, determines the optimal investment plan for adding power plants over a planning horizon to meet the demand criteria, fuel mix ratio, and the reliability criteria. To test the proposed SFLA method, it is simulated for two test systems in a time horizon of 10 and 20 years respectively. The obtained results show that compared to the traditional methods, the SFLA method can provide better solutions for the GEP problem, especially for a longer time horizon.