Robust possibilistic programming for multi-item EOQ model with defective supply batches: Whale Optimization and Water Cycle Algorithms
Robust possibilistic programming for multi-item EOQ model with defective supply batches: Whale Optimization and Water Cycle Algorithms
复制标题
DOI：
10.1007/s00521-018-3492-3
复制
发表时间：
2019-10
影响因子：
6
通讯作者：
Soheyl Khalilpourazari;S. Pasandideh;A. Ghodratnama
中科院分区：
文献类型：
--
作者：
Soheyl Khalilpourazari;S. Pasandideh;A. Ghodratnama
针对非理想供应批量的多产品经济订货问题，提出了一种新的数学模型。供货批次在到货时按“全有或全无”的政策进行检验，如果发现有缺陷，整批将被拒收。本文的目标是确定每种产品的最优订货量和缺货规模。为了建立一个现实的数学模型，提出了三种稳健的可能性规划(RPP)方法来处理模型主要参数的不确定性。针对已提出的RPP模型的复杂性，提出了两种新的元启发式算法--水循环算法和鲸鱼优化算法来求解RPP模型。解决了各种测试问题，以评估两种新型元启发式算法使用不同度量的性能。通过单因素方差分析和Tukey的HSD检验，比较了两种元启发式算法的有效性。比较了RPP模型与基本可能性机会约束规划(BPCCP)模型在不同实现方式下的适用性和效率。仿真结果表明，RPP模型的性能明显好于BPCCP模型。最后，进行灵敏度分析，以确定数学模型主要参数的任何变化对目标函数值的影响，从而确定最关键的参数。
This paper proposes a new mathematical model for multi-product economic order quantity model with imperfect supply batches. The supply batch is inspected upon arrival using “all or None” policy and if found defective, the whole batch will be rejected. In this paper, the goal is to determine optimal order quantity and backordering size for each product. To develop a realistic mathematical model of the problem, three robust possibilistic programming (RPP) approaches are developed to deal with uncertainty in main parameters of the model. Due to the complexity of the proposed RPP models, two novel meta-heuristic algorithms named water cycle and whale optimization algorithms are proposed to solve the RPP models. Various test problems are solved to evaluate the performance of the two novel meta-heuristic algorithms using different measures. Also, single-factor ANOVA and Tukey’s HSD test are utilized to compare the effectiveness of the two meta-heuristic algorithms. Applicability and efficiency of the RPP models are compared to the Basic Possibilistic Chance Constrained Programming (BPCCP) model within different realizations. The simulation results revealed that the RPP models perform significantly better than the BPCCP model. At the end, sensitivity analyses are carried out to determine the effect of any change in the main parameters of the mathematical model on the objective function value to determine the most critical parameters.