Capacitated lot sizing problem using adaptive multi-chromosome crossover strategy

The capacitated lot sizing problem (CLSP) involves determining optimal lot sizes for different products subject to certain capacity constraints. Mathematical programming works effectively to solve small scale problems, but, as the problem size increases, computation using such techniques becomes impractical. Genetic algorithms (GA) are one of the meta-heuristics commonly used to solve such problems. The multi chromosome crossover strategy (MCS) has been applied previously in GA to solve the CLSP and found better results than with classical crossover. This paper evaluates the effectiveness of using three different crossover strategies - classical, MCS and adaptive MCS. The results indicate that the adaptive strategy generates the optimal solution more frequently than the previous two strategies.