Supply chain network optimization model incorporating competitive facility location problems

The detailed literature review on SC network models showed that almost all of the SC network optimization models ignore the impacts of SC network decisions on the customer demand. However, the physical network structure of a SC is one of the important factors impacting chain's competitiveness, especially for the retail markets. On the other hand, competitive facility location problems model only distribution part of the SC even though they have some characteristics of SC networks and analyze the rival chains existing in the market. In this dissertation study, a multi-objective SC network optimization model, incorporating competitive facility location models is developed. The objectives utilized in the model are; profit maximization, sales maximization and SC risk minimization. The model is defined as Mixed Integer Linear Programming (MILP) model with three echelon SC network, with multi products, and single term. The SC structure consists of three echelons; Suppliers, Distribution Centers and Customer Zones. In order to simplify the model, products are aggregated to 10 different product types to represent the whole product mix. Nature of the developed model is deterministic. The unique unknown variable within the model is the demand. The demand at each customer zone is assumed to be determined by the price and the attraction function. Attraction Utility function is defined as the availability of same-day transportation from DC to Customer Zone. The model is applied to a real life problem of one of the leading ready – wear clothing companies in Turkey. The company currently has only one DC in Istanbul. However, the number of sales points and the total sales volume of the company increased sharply, in recent years. It is considered that the firm needs to reconfigure its supply chain network and to decide whether or not to open additional Distribution Center(s) in alternative locations. The model is first solved as a profit maximization problem. In the optimal solution for profit maximization problem, the sales decreased by 26 %. Because of the competition within the market, the 26 % sales decrease is not acceptable by any firms. Therefore, it has been concluded that modeling the problem as profit maximization does not generate required results and all three performance measures (profit, sales, risks) need to be taken into account. Then, the model is solved as multi-objective with a goal programming method. The results proved that including attraction function on the model may change the performance results of the model and eventually SC network decisions. When the model moves from one DC (current situation) to two DCs, the model generates around 5 % more sales volume due to the defined attraction function. Without including attraction function impact on the model, one DC options are chosen. However, the model with the attraction function proposes DMs to utilize two DCs concurrently. The result of this study contributes to the Supply Chain Network Optimization model literature mainly in two ways. First, the developed and analyzed model is the first SC network optimization model incorporating the changes in the demand, which is subject to the both price and distance from the end-customers and which is substantially influenced by strategic level SC network model decisions. Second, this model is also the first model simultaneously utilizing supply side risk analysis, demand functions and strategic level SC decisions. Developed model also proved that single objective models may not generate acceptable results and showed that SC network optimization models need to be defined as multi-objective since SCs are multi-objective in their nature.