An Improved Genetic Algorithm for Solving Tri-level Programming Problems

When genetic algorithm is adopted to solve tri-level programming, many problems exist, such as controlling population size, jumping out of local optima, and avoiding low efficiency. An improved genetic algorithm with parallel strategy is proposed in this paper to solve tri-level programming, as well as elites reserving and fitness value crowding strategy. Simulations with numerical examples are done to prove correctness and effectiveness of the proposed algorithm.