|Abstract: ||在今日競爭激烈的全球商業環境下，產品日益翻新及生命週期較過去為短，因此，適時與正確的生產規劃(production planning)已成為一個管理者最重要的決策問題。然而，生產規劃本質上是一種複雜的多目標決策(multiple objective decision making,MODM)，決策過程中所需的資訊通常存在不確定(uncertain)、含糊(vague)或不明確(imprecise)的資料。在文獻上，多目標決策主要應用在「規劃面/設計面」，通常探討在不同的限制條件下追求多個目標的同時達成，且獲得一個有效解(efficient solutions)或非劣解(non-inferior solutions) 之解集合。去年本計畫著重於解模糊二階多目標生產規劃模式，該模式是運用Sakawa方法求取二階廠商雙方最大滿意水準。今年本計畫修正去年研究方法應用於三階( 含) 以上之模糊多階多目標生產規劃模式(fuzzy multi-levelmulti-objective production planning models)，再利用基因演算法求解並比較兩者的求解績效。
multi-objective production planning, genetic algorithm Nowadays, due to the drastic competition in a global business environment, products are renovated quickly and their life cycles are shorter than before. To prepare an effective production plan is the most crucial decision problem for a production manager. In fact, production planning is essentially a complicated multiple objective decision making (MODM). The information required in the decision process usually involves uncertain, vague or imprecise data. In the literature, the MODM is applied to some aspects, including planning and design. It aims at achieving the best levels of relevant objectives simultaneously subject to some constraints, and getting a set of efficient solutions (i.e., non-inferior solutions). In the last year, this study focused on solving fuzzy two-level production planning models by using Sakawa’s method to obtain the maximized satisfaction level. In this year, however, we will revise Sakawa’s method, propose a genetic algorithm, and then compare the performance of the two methods in solving fuzzy multi-level multi-objective production planning models. In the past, most related researches were concentrated on multiple objectives (e.g., maximization of profit, minimization of cost, minimization of change rate of workforce level, and maximization of service level) set only by the manufacturers. The interdependent relationship between a wholesaler (or a retailer or a customer) and a manufacturer as well as that between a manufacturer and a supplier had been rarely discussed. This study will consider these relationships in terms of a manufacturer’s fuzzy production planning to incorporate its upstream and downstream members’ objectives and constraints into the manufacturer’s production planning decision process so that the benefits of all members in the supply chain can be attained. In this study, the results of the first year’s research will be described at first and then several fuzzy multi-level multi-objective functions and their related constraints will be built. Next, for a multi-level supply chain, we will revise Sakawa’s method, and propose an efficient genetic algorithm, and analyze the performance of the two proposed methods through conducting an experimental design. Finally, the proposed genetic algorithm will be applied to a well-known Taiwanese network equipment manufacturing supply chain. The conclusions drawn from this case study will be referred by practitioners.
This study can be divided into the following six parts:
1. Review the related literature.
2. Describe the research problem.
3. Build fuzzy multi-level multi-objective production planning functions and their related constraints.
4. Introduce Sakawa’s method and existing approaches for solving fuzzy multi-level multi-objective planning models and explain the reason why Sakawa’s method will be revised.
5. Develop an efficient genetic algorithm and a revised Sakawa’s method, design an experiment, and compare the performance of the two proposed methods.
6. Study a practical case of a well-known Taiwanese network equipment manufacturing supply chain.