A. Rajabian, S. M. H. Hosseini, H. Amouzad Khalili, M. Amirkhan,
Volume 21, Issue 3 (9-2024)
Abstract
Mathematical modeling in multi-echelon and multi-product sustainable biomass supply chain design investigates the behavior and relationship between the factors affecting this network. These include different network dimensions, the amount of demand and supply, resource requirements, restrictions, and limitations. Various methods such as simulation and optimization are usually used for mathematical modeling of the multi-level and multi-product sustainable biomass supply chain network. By using these methods, with the help of existing optimization algorithms, it is possible to accurately and quickly predict the necessary costs for the implementation of the network. Also, mathematical modeling in the field of multi-level and multi-product sustainable biomass supply chain network, taking into account economic, social, and ecological aspects at the same time, allows us to use appropriate algorithms to optimize the necessary factors for consider the implementation of the network and look for ways to reduce costs while maintaining the quality of services and increasing social welfare. In this research, a mixed integer multi-objective mathematical programming model for use in a multi-level and multi-product biomass supply chain is presented. The proposed model has been evaluated using the NSGA-II meta-heuristic algorithm, which can determine the optimal values for the main components of the supply chain in the long term. The most important values that we can calculate using this model are the optimal amount of production of glycerin, biodiesel, Jatropha, algae, and the amount of wco in biological refineries, determining the optimal amount of Jatropha oil and algae, determining the amount of production of drug raw materials and fertilizer production in Extraction centers, Jatropha, Algae, Biodiesel, Waste Cooking Oil (WCO), Norozak, Fertilizer, and Dates, Determining the capacity of oil collection and extraction centers. In addition, a sensitivity analysis has been performed with changes in the epsilon value. According to the obtained results, the level of significant changes of epsilon between 50 and 900 is determined as the improving vector. Therefore, the range of epsilon changes to search for the local optimal solution for the first objective function is set at 650 and for the second objective function at epsilon 600. In addition, changes in the value of the objective functions have been calculated in case of changes in the capacity of the fields.
A. Yaghoubi, A. Rajabi,
Volume 22, Issue 2 (6-2025)
Abstract
Among the important and effective responsibilities in industrial units is planning and inventory control. In this case, determining the optimal balance between inventory levels, ordering, holding, and purchasing costs plays a significant role in preventing capital waste and dealing with inventory shortages. In this regard, this paper designs an economic production quantity (EPQ) model in single and multi-product cases with considering production capacity constraints, general discount on purchase orders for items and price dependence on the order quantity. This model is applicable when the production rate of products is lower than their demand rate and the production system faces production capacity constraints. Also, it is assumed that the cost function of purchasing each unit of goods follows an exponential distribution and depends on the quantity of purchase orders. The presented model cannot be solved in polynomial time, and as the number of products increases, the problem solving time increases with an exponential distribution. So, in order to solve the model in the single-product case, the GAMS software was used, and for multi-product cases, the Genetic Algorithm (GA) and Imperialist competitive Algorithm (ICA) were used. In order to improve the performance of the solution algorithms, response surface methodology (RSM) has been used to determine the optimal values of their parameters. Finally, in order to evaluate the performance of the proposed model, random problems were designed and the results were evaluated by the indicators of solution quality and run time, which indicate the superiority of the GA.