本研究主要探討大型量販店物流網路之區位定址問題，並加入供應商管理庫存策略，並整合了供應鏈供應商選擇、庫存控制與運輸等供應鏈中重要議題，以兩階層的供應鏈為例，由供應商將商品成品運送至配銷中心，配銷中心再將商品成品送至消費者手中。 由於本模式所考量的目標為總成本最小、服務回應率最大、商品訂單達交率等三大目標，並非單一的目標，所以我們使用了基因遺傳演算法(NSGA-II)求解混合非線性整數規劃問題，並採行兩個貪婪法則來使用基因遺傳演算法求取柏拉圖最佳解。 依據所建立問題之混合式基因演算法所寫的MATLAB程式進行相關的進行模式求解與數據分析，以便能瞭解數學規劃模式中相關參數與不同因子的變化，對於整合性物流網路選址問題當中之最適解所產生的影響性。 最後，我們設計了不同情境，將不同目標函數賦予不同的權重值，以得知對於不同目標函數權重的情境明顯會影響配銷中心的地點設置之決策，再由決策者依據所得的結果選擇適合的方案，最後再把成本項細分為六項，探討各項成本的影響程度。 The study focused on the Facility Location Problem (FLP) problem in a logistics network for a large retail store in Taiwan. We also add in the VMI strategy and also provide the solutions for the supplier selection, inventory control, and transportation decisions. Using a two-stage supply chain with as an example, the suppliers send product to distribution centers (DCs) and DCs send products to consumers.Because the goal of our three objective are to make the total cost smallest, Responsiveness Level biggest and volume fill rate biggest, not single objective, so we use NSGA-II to solve mixed nonlinear integer programming problem and take two greedy approach to use heuristic genetic algorithm and find the optimal solution of Pareto. According to the established model and we used the software of MATLAB to solve and analyzed the data, then we can understand the change of our model’s parameters and different factors, than we can realize the effects of the optimal Pareto solutions. Finally, we designed several different conditions and distributed different weight to different objective, so we can know different condition would obviously affect the strategy of facility location, than the decision maker could accord to the result and found a solution. After we got the result, we divided the total cost into six items of cost, and tried to know the impact of the six different items.
