Scholarlink

<p>One of the critical steps in the minimization of time and costs and resolving environmental concerns in closed-loop supply chain networks is the modeling and optimization stage. This is done through the evaluation of the alternative scenarios based on the locations within the chain and the available resources. The same fact motivates the use of closed-loop supply chains for the purpose of the adoption of a sustainable and efficient strategy in a variety of industries. Thus, the motivation of writing the present paper is to address a multi-objective, multi-echelon, closed-loop supply chain in order so as to solve real-world challenges practically. Disposal and recycling centers and also revival and remanufacturing centers are utilized in the present problem in order to decrease costs and secure environmental dimensions. The supply chain considered in the present research was studied under unknown demand with regard to service level in order to optimize four major objectives simultaneously. The objective functions include (I) decreasing the total expected cost, (II) enhancing the service level, (III) the minimization of emitted greenhouse gas (GHG), and (IV) enhancing job creation and employment. The primary novelty of the present paper is to forecast unknown demand by employing machine learning approaches with regard to customer service level. As a result, by utilizing a mixed-integer linear programming (MIP) model, this paper presents a multi-objective optimization framework. Furthermore, this investigation presents a machine learning-based prediction process in order to determine the customers’ demands. One may use the suggested solving process to resolve the problems by employing the data obtained from a real-life case investigation with regard to the limitations and requisite challenges. According to the results, the suggested machine learning process may accurately anticipate the demand, and the solution technique may provide managers with appropriate alternatives with regard to the preferences of different stakeholders. Furthermore, supplementary analysis presented suitable trade-offs between the managerial objectives to select the optimum solution on the basis of their preferences.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Multi-Objective Optimization Framework to Design the Closed-Loop Supply Chain Network Using Machine Learning for Demand Prediction

  • 期刊论文

  • 发表日期：

    2025年5月2日

查看全文
  • Omid Rezaei,
  • Rashed Sahraeian,
  • Seyed Mohammad Hassan Hosseini

摘要

One of the critical steps in the minimization of time and costs and resolving environmental concerns in closed-loop supply chain networks is the modeling and optimization stage. This is done through the evaluation of the alternative scenarios based on the locations within the chain and the available resources. The same fact motivates the use of closed-loop supply chains for the purpose of the adoption of a sustainable and efficient strategy in a variety of industries. Thus, the motivation of writing the present paper is to address a multi-objective, multi-echelon, closed-loop supply chain in order so as to solve real-world challenges practically. Disposal and recycling centers and also revival and remanufacturing centers are utilized in the present problem in order to decrease costs and secure environmental dimensions. The supply chain considered in the present research was studied under unknown demand with regard to service level in order to optimize four major objectives simultaneously. The objective functions include (I) decreasing the total expected cost, (II) enhancing the service level, (III) the minimization of emitted greenhouse gas (GHG), and (IV) enhancing job creation and employment. The primary novelty of the present paper is to forecast unknown demand by employing machine learning approaches with regard to customer service level. As a result, by utilizing a mixed-integer linear programming (MIP) model, this paper presents a multi-objective optimization framework. Furthermore, this investigation presents a machine learning-based prediction process in order to determine the customers’ demands. One may use the suggested solving process to resolve the problems by employing the data obtained from a real-life case investigation with regard to the limitations and requisite challenges. According to the results, the suggested machine learning process may accurately anticipate the demand, and the solution technique may provide managers with appropriate alternatives with regard to the preferences of different stakeholders. Furthermore, supplementary analysis presented suitable trade-offs between the managerial objectives to select the optimum solution on the basis of their preferences.