A New Hybrid Genetic Algorithm for the Stochastic Loader Problem

In 2004, Tang proposed a new NP-hard combinational optimization problem that frequently arises in practice - The Loader Problem. Two special cases of the problem (the restricted loader problem and the equal loader problem) and optimal solution strategy have been considered. In this paper, we extend Tang´s model by proposing the stochastic quantity of load and unload at each station that make the model more applicable in practice. For finding the optimal solutions, we present a new hybrid genetic algorithm that combines self-adapting crossover and stochastic mutation operators. Comparing with the basic genetic algorithm, this improved algorithm adequately utilizes the adaptability information of current individuals and has better convergence efficiency and higher solution precision. Two numerical examples illustrate the validity and efficiency of the new hybrid genetic algorithm.