Selective Control of Independent Activities: Linear Programming of Markovian Decisions

This study is related to optimum control strategies used by a manager supervising a group of independent activities whose performances deteriorate with time. At the end of a day, the manager evaluates information regarding the performance levels of the activities during that day and decides which activity to attend to, since he can control only one at a time. The deterioration of performance and the improvement of the deteriorated performance by a control action from one level to other levels are given by stationary transition probabilities. Optimal strategies for this control are obtained by the linear programming method proposed by Manne [Manne, A. S., Linear programming and sequential decisions. Management Science Vol. 6 No. 1 (April 1960), pp. 259–267.].