Supply chain management is a burning issue for modern industrial enterprises. To handle this issue, non-linear stochastic models are successfully applied to find the reasonable and efficient solutions. A need to develop a unique method to find the solutions to supply chain management tasks defined as stochastic mixed-integer non-linear programming tasks is determined by the limitations imposed by the general models. The sum of the total raw procurement costs from the Commodity Exchange over the defined planning horizon is taken to be the target function of the unique model, while the binary variables which show whether a purchasing order is included into the procurement plan are used for optimization purposes. Some parameters of model’s limitations are stochastic and consider the uncertainty factor and risks in supplying the required raw materials to the manufacturing site. Branch-and-bound and genetic algorithms are applied at some steps in the developed heuristic algorithm. The algorithm and the model are tested at a major timber processing enterprise in Primorsky Area. Four types of processors over three planning horizons were applied to compare the efficiency of the proposed algorithm with partial application of the genetic algorithm or branch-and-bound method. The findings analysis shows that, unlike the genetic algorithm, the unique one is more stable in terms of uncertainty of the input parameters in comparison with the branch-and-bound method. It provides the solutions in the models with a great number of variables. The algorithm is shown to be universal enough for its further modification in solving more complicated problems of the same class, containing a significantly larger number of probabilistic parameters that describe other uncertainties in the supply of raw materials. Further research is seen to include the development of the proposed algorithm to increase the rate of convergence for its better efficiency.
A Hybrid Genetic Algorithm-Simulated Annealing for Integrated Production-Distribution Scheduling in Supply Chain Management
