Hybrid optimization algorithm for scheduling decision support

Genetic algorithms are stochastic methods for solving search and optimization problems. Simulated annealing is another stochastic method for finding optimal values numerically without trapping to local minimum or maximum. In this paper a hybrid algorithm that combines the benefits of the both algorithms is presented. The implementation of the hybrid algorithm is presented and tested in the thermo-mechanical pulp (TMP) production scheduling, which is a dynamic, combinatorial optimization problem. Due to a high electricity consumption in TMP production, the cost savings of optimal scheduling are up to millions of €/a at one production site. The results show that the hybrid algorithm is an improvement when compared to the plain algorithms. However, choosing appropriate parameter settings for the method is a demanding task and essential to the efficiency of the algorithm.