Scholarlink

<p>Efficient scheduling of unrelated parallel machines with energy considerations and maintenance constraints is crucial for reducing operational costs, improving sustainability, and enhancing productivity in modern manufacturing systems. Indeed, machines consume varying amounts of energy depending on their operational state, with idle machines requiring less energy than active ones. Additionally, energy consumption fluctuates based on the type of production job and the specific machine used. The optimization problem addressed in this study involves both time and energy considerations, aiming to optimize time-based scheduling efficiency and energy consumption. More formally, it aims to minimize two objectives: the first seeks to reduce earliness and tardiness in production and maintenance, while the second focuses on minimizing energy consumption. To achieve better decision-making by not prioritizing one criterion over another, but rather selecting a balanced solutions, a bi-objective resolution approach is employed. A Mixed Integer Linear Program (MILP) is formulated, utilizing the <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\epsilon \)</EquationSource> <EquationSource Format="MATHML"><math> <mi>ϵ</mi> </math></EquationSource> </InlineEquation>-constraint method to compute the exact Pareto front for small-sized instances. For larger instances, an adaptation of the Multi-Objective Simulated Annealing algorithm (MOSA) is introduced to approximate the Pareto front. In this enhanced version of MOSA, referred to as POP-MOSA, a population-based approach is integrated, along with strategies to refine the search process. The proposed methods are validated using two types of data: theoretical data generated from existing literature and real-world data collected from the PROBOT educational platform at the University of Technology of Troyes (UTT). Since the PROBOT platform consists of only two robots, the dataset is extended to create relatively large instances. MILP model successfully computes the complete Pareto front for up to 30 production jobs and 2 machines within an execution time of less than one hour. Furthermore, various performance metrics demonstrate the effectiveness of the proposed POP-MOSA approach.</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

Green unrelated parallel machine scheduling problem under availability and energy constraints: a bi-objective approach

  • 期刊论文

  • 发表日期：

    2026年4月1日

查看全文
  • Meriem Touat,
  • Karima Benatchba,
  • MohammadMohsen Aghelinejad,
  • Lyna-Razane Meguellati

摘要

Efficient scheduling of unrelated parallel machines with energy considerations and maintenance constraints is crucial for reducing operational costs, improving sustainability, and enhancing productivity in modern manufacturing systems. Indeed, machines consume varying amounts of energy depending on their operational state, with idle machines requiring less energy than active ones. Additionally, energy consumption fluctuates based on the type of production job and the specific machine used. The optimization problem addressed in this study involves both time and energy considerations, aiming to optimize time-based scheduling efficiency and energy consumption. More formally, it aims to minimize two objectives: the first seeks to reduce earliness and tardiness in production and maintenance, while the second focuses on minimizing energy consumption. To achieve better decision-making by not prioritizing one criterion over another, but rather selecting a balanced solutions, a bi-objective resolution approach is employed. A Mixed Integer Linear Program (MILP) is formulated, utilizing the \(\epsilon \) ϵ -constraint method to compute the exact Pareto front for small-sized instances. For larger instances, an adaptation of the Multi-Objective Simulated Annealing algorithm (MOSA) is introduced to approximate the Pareto front. In this enhanced version of MOSA, referred to as POP-MOSA, a population-based approach is integrated, along with strategies to refine the search process. The proposed methods are validated using two types of data: theoretical data generated from existing literature and real-world data collected from the PROBOT educational platform at the University of Technology of Troyes (UTT). Since the PROBOT platform consists of only two robots, the dataset is extended to create relatively large instances. MILP model successfully computes the complete Pareto front for up to 30 production jobs and 2 machines within an execution time of less than one hour. Furthermore, various performance metrics demonstrate the effectiveness of the proposed POP-MOSA approach.