RESUMO
Facility location, particularly in the context of international investments by global enterprises, stands out as a paramount concern within the purview of top management's strategic decision-making process. The selection of a suitable location plays a pivotal role in determining the ultimate achievement of organizational objectives. The process of selecting an appropriate location requires the comprehensive analysis of a substantial volume of data, encompassing diverse tangible and intangible evaluation criteria that may exhibit inherent conflicts. This paper addresses the challenge of determining the best location for a manufacturing facility by employing alternative performance measures within the framework of the data envelopment analysis (DEA) model. In a performance evaluation process, not only positive but also negative aspects should be determined. This paper, therefore, proposes a double-frontier DEA-AR model, which is an integrated approach that incorporates the efficient frontier, anti-efficient frontier, and assurance region weight restrictions, with the aim of increasing the discrimination ability of the DEA method. An efficient frontier evaluates the information of each location from a positive viewpoint, while the worst side is evaluated by an anti-efficient frontier. The technique of weight restrictions, which allows incorporating expert opinion into the assessment, is also applied with both frontiers to restrict the regions of weights to some specific area. The prescribed approach is illustrated by a numerical example of selecting the best location among ten different countries under consideration of 22 selection criteria obtained from PEST analysis. The results show that the proposed alternative performance measures significantly improve discrimination capability, enabling the ranking of candidates based on their suitability for the optimal location.
RESUMO
Pork stands out as the most extensively produced and consumed meat globally. With advancements in technology, genetics, and management, the structure of the pig supply chain has transformed from the traditional birth-to-slaughter raising method to incorporate four primary specialized operations: breeding, farrowing, nursery, and fattening. Fattening, constituting approximately 70% of a market pig's entire life cycle, heavily relies on resources, notably in feed consumption. Despite the integration of feed production with pig farming in modern industrial setups through farming contracts, separate decision-making processes for production planning in both stages often result in overall inefficiency. This research proposes an optimization-based methodology to plan production for a vertically integrated setting of three supply chain echelons: a feed mill, fattening farms, and a slaughterhouse. Key coordinated decisions include creating production plans for specific feed formulations at the feed mill and organizing farming cycles at fattening farms to meet the demand of the slaughterhouse The aim is to optimize pig growth while minimizing the overall costs. The methodology includes a mixed-integer linear programming model for the pig supply chain, and a Lagrangian heuristic as method to make coordinated production plans. Computational experiments were conducted using diverse case-study data based on pig supply chains in Thatland. Compared with the results using a commercial software, Lingo's Simplex method, our proposed heuristic could find optimal solutions quicker for smaller problem instances and produce more effective feasible solutions within limited time frames for larger scenarios.
