An Efficient Computational Alternative to 'Using Linear Programming to Design Oil Pollution Detection Schedules'.

摘要

In Olson, Wright, and McKell's recent paper on the design of oil pollution detection schedules, an interesting and inventive development and application of a Markov Decision Process was presented. Optimal schedules for patrol flights of surveillance aircraft were found using linear programming. In this paper the model has been reformulated as a discrete time semi-Markov process. Significant computational advantages accrue from this alternative approach. (Author)