مشکل تعیین اندازه دسته تولید تصادفی پویا با استفاده از برنامه نویسی دو سطحی بر اساس تکنیک های هوش مصنوعی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|22800||2012||14 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Applied Mathematical Modelling, Volume 36, Issue 5, May 2012, Pages 2003–2016
Simulation is generally used to study non-deterministic problems in industry. When a simulation process finds the solution to an NP-hard problem, its efficiency is lowered, and computational costs increase. This paper proposes a stochastic dynamic lot-sizing problem with asymmetric deteriorating commodity, in which the optimal unit cost of material and unit holding cost would be determined. This problem covers a sub-problem of replenishment planning, which is NP-hard in the computational complexity theory. Therefore, this paper applies a decision system, based on an artificial neural network (ANN) and modified ant colony optimization (ACO) to solve this stochastic dynamic lot-sizing problem. In the methodology, ANN is used to learn the simulation results, followed by the application of a real-valued modified ACO algorithm to find the optimal decision variables. The test results show that the intelligent system is applicable to the proposed problem, and its performance is better than response surface methodology.
Determining proper production and replenishment quantity is one of the critical and active research topics of supply chains. To increase competitiveness and efficiency, upstream and downstream members of the packing industry usually integrate as a supply chain. Because of weather conditions, the acquired quantity of industrial raw materials tends to be non-deterministic and uncontrollable during the farming process. Such raw materials also deteriorate easily. Therefore, in this complex environment, decision makers need a method to determine a proper budget and reduce the raw material deterioration. In previous studies, replenishment problems seldom considered non-deterministic raw material and demand. Vidal and Goetschalckx  reviewed some important works about production and distribution models. This paper adopted artificial intelligence techniques to solve a stochastic replenishment problem. This paper proposes a bi-level programming problem. The upper level problem is stochastic. This section determines the budget to reduce raw materials’ deterioration rate. The lower level problem is a replenishment problem during simulation. Ben-Ayed et al.  introduced the characteristics of the bi-level programming problem. Kalashnikov and Ríos-Mercado  proposed a direct algorithm to solve mixed integer bi-level linear programming. Their problem is a deterministic model. Campêlo and Scheimberg  adopted the simplex method to solve a deterministic linear bi-level program. Mishra and Ghosh  proposed an interactive fuzzy programming method to solve bi-level quadratic fractional programming problems. Colson et al.  proposed an approximation algorithm to discover a near-optimal solution for nonlinear bi-level programming. The aforementioned works utilized local optimization methods, and their problems are all deterministic models. This paper uses a novel method to solve a stochastic replenishment problem. The proposed problem obtains expected cost by simulation model, and the simulation process is an NP-hard problem. Main contributions of this paper include: (1) proposing a bi-level programming model for NP-hard stochastic replenishment problem, (2) adopting a novel method and solution to proposed problem, (3) providing an illustrative example to demonstrate the methodology, and (4) notices for using the method under numerical analysis. This paper proposes a stochastic dynamic lot-sizing problem with asymmetric deteriorating commodity (SDLSPADC). Artificial intelligence techniques are becoming more and more advanced and widely implemented. This paper adopts an intelligent decision system based on artificial neural network, data mining and ant colony optimization (neuro-DM&ACO) to solve the proposed bi-level programming. During the simulation process, this simulation-based optimization system includes an NP-hard replenishment problem (i.e., lower level problem). The organization of this paper is as follows: Section 2 formulates SDLSPADC. Section 3 constructs the neuro-DM&ACO model to solve SDLSPADC. Section 4 compares performances of neuro-DM&ACO and response surface methodology. Section 5 summarizes the results and presents recommendations for further research.
نتیجه گیری انگلیسی
Simulation technique is often used on analyzing stochastic inventory problems. The paper proposes a stochastic bi-level programming problem. The lower level problem is NP-hard during simulation procedure. This paper utilizes an artificial intelligence method and neuro-DM&ACO to solve a stochastic replenishment problem. This proposed novel methodology can handle several obstacles in traditional simulation techniques: (1) solutions for replenishment policy problem in which the simulation process includes NP-hard problem, (2) effectively discovering data for constructing ANN simulation model through RSM, and (3) providing a methodology for stochastic bi-level programming model that is easy to operate. Performing improper BPN parameters in the neuro-DM&ACO approach causes network training into local optimum. To optimize RSM decision variables, the integration of RSM and ACO successfully finds the stationary point within the minimum response. The solution of the neuro-DM&ACO approach is better than that of RSM in our confirmation experiment. The manager can refer to the results of this experiment when determining the replenishment policies in similar supply chain environments. This paper suggests two directions for future research: (1) the simulation model of this paper is constructed with weight and threshold value of the ANN (i.e., Table 7), and other heuristic algorithms are compared with the optimization method in this paper, and (2) this paper illustratively uses novel technique for stochastic bi-level programming problem. Future research may test the impact of the method or sensitivity analysis under different scenarios.