نوع مقاله : پژوهشی
نویسندگان
گروه مهندسی صنایع، دانشکده فنی مهندسی گرگان، دانشگاه گلستان
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسنده [English]
In this paper, a mathematical model has been developed with the aim of locating branch warehouses of a pharmaceutical distribution company, taking into account the existing legal requirements regarding drug distribution. In this model, in addition to the structure of the distribution network, the optimal flow of contracted drugs at the level of the company's branches throughout the country is also determined. Due to the high complexity and high dimensions of the model in real conditions, a combined solution approach based on meta-heuristic methods has been proposed. To solve the model, at the beginning, the structure of the model was decomposed into a main problem and several sub-problems and then it was solved using a two-stage genetic algorithm. In the first stage, it solves the main problem and in the second stage, it solves sub-problems. In order to apply the balance constraints related to the connection between stages in the supply chain (input or output flow of or from each stage to others), priority-based encoding was utilized in the second level. Since in solving the model with real dimensions, due to the high dimensions of the problem in addition to its complexity, solution time is very high, the p-medoid clustering method was incorporated to aggregate supply chain customers. Finally, the tuning of algorithms was used with the popular Taguchi method. In order to validate the model, a case study using real data from Elite Daru Distribution Company has been considered. Results of the case study indicate a 23% reduction
in distribution costs in optimal design as compared to existing design. Studies have also shown that we face a 7% cost in the case of restrictions imposed by law compared to the case without it. Sensitivity analysis on the number of cluster centers showed that data integration would lead to an average cost change of less than 1%.
کلیدواژهها [English]