Search

A systematic process to help the buyer decide on what particular fashion goods items to choose for a catalogue merchandiser is presented. The problem is that delivery lead times are long while the public's fashion sense and the general economic condition can change radically. It has been the practice of a buyer to rely mainly on his/her judgment. The buyer has to decide how much to order and later how much to transfer to the outlet stores depending on actual demand. A model of demand for an individual item was developed based on historical data and buyer judgment. A stochastic dynamic programming (DP) model of the buying problem which considers the model of demand was then formulated. The DP model was then employed to determine the structure of the optimal inventory policy which was found to be substantially superior to the current method.

During the past year the facts of life for business have come to include such unpleasant matters as material shortages and unstable exchange rates. These have focused new attention on the strategic factors of the supply function of many companies, both domestic and international. In this article, the authors isolate and discuss seven factors derived from a survey of 110 companies and interviews with purchasing executives. These factors range all the way from distance to the effects of nationalism. The impact of each factor on sourcing strategy is considered by itself. Thus the authors identify the effect of each for the international company, for the company with only local operations, and for the one with one or two "venture-minded" operations at some distance from the bulk of its units. Then, in view of each of these factors, they examine the trade-off between centralized and decentralized purchasing functions, and the question of how global a search for suppliers should be. [ABSTRACT FROM AUTHOR]

Service and manufacturing companies, particularly the automobile industry, typically make production capacity decisions before knowing the level of consumer demand. As a result, companies maintain capacity in excess of optimal levels in relation to actual output. A decision-making model using Lotus 1-2-3, which incorporates scenario planning, integer programming, and risk analysis, was developed for General Motors (GM) to aid in determining capacity for four of their automobile lines and to help management understand the tradeoffs between profit and risk in capital planning. The model confirmed management intuition that GM had excess capacity and that, to minimize costs, several plants had to be closed. The model recommended a specific subset of products not in current product lines and supplied cost information about new configurations that management had developed to optimize production capacity.

Mixed-integer programming models are generally not used to solve real-world production scheduling problems because these models have exorbitant solution times. Alternative formulations are presented that significantly improve model improvement times. Mixed-integer programming models can be used to near optimality when used in conjunction with variable redefinition methods. One significant aspect of the research is that it allows practitioners to avoid the complications of Lagrangian relaxation or column generation methods; the variable redefinition results can be applied using 'off-the-shelf' codes such as LINDO or MPSX-370.