신규지점의 운영설계 및 생산목표설정: 다목적최적화기법과 DEA의 결합사용

This study explores an important managerial problem of operational design and production target setting for opening branches in a company. There are various cases of opening new branches in many different kinds of industries. Examples are fast-food franchise restaurants, telecommunication service offices, and bank branches. Generally, these branches each utilize multiple inputs, such as manpower, operating costs, and facilities, to produce multiple outputs like revenue and customer satisfaction. To operate such a new branch, it is therefore needed to plan as to how much or many each of inputs should be and how much or many each of outputs should be produced, among others. In this article, we refer to input setting for a new branch as operational design, and output setting as production target setting. The primary purpose of this study is to develop a unified procedure for optimal operational design and production target setting. We propose a methodology that jointly uses data envelopment analysis (DEA) and multi-objective optimization (MOO) techniques, as well as statistical techniques. To show the usability and applicability of the developed method, we apply it to a fast-food company (FFCOM) in Korea which has over 180 branches on a nationwide scale. All branches perform common tasks for providing fast-food services with customers, using multiple inputs to produce multiple outputs. Whenever new branches open, they decide the levels of operational design and production target by their own experience or intuitive knowledge, without a systematic or scientific approach to the problem in question. There is therefore no guarantee that the solution becomes an efficient or optimal one. Inefficiency implies an excessive use of inputs and/or shortfall of outputs to be produced by a branch relative to the other branches of interest. Thus the design and target solution needs to be efficient in comparison with some or all of the experimental input-output data. DEA is a mathematical programming approach to the evaluation of relative efficiency of branches. It clearly identifies the efficient branches. For reasons like these, we employ DEA, as part of the procedure we propose, to guarantee efficient solution. Meanwhile, there are many efficient combinations of inputs and outputs in general. An optimal or most preferred solution needs to be determined, among the efficient combinations, for use in operating the new branch. Basically, the optimal solution maximizes output functions and minimizes input variables, simultaneously. This leads us to the consideration of a MOO formulation and then use of MOO solution techniques in order to arrive at an optimal design and target solution. We stress the usefulness of interactive MOO approach at this final stage of the procedure.