A recursive algorithm for volume-based sensitivity analysis of linear decision models

One problem impeding the use of linear decision models in practical applications is the difficulty in precisely specifying weights. In this paper, we analyze an approach to establish the stability of a decision, given approximate information on weights. This approach is based on comparing the volumes of regions in weight space, in which different alternatives are optimal. To compute those volumes, a recursive algorithm was developed. The efficiency of that algorithm is analyzed both analytically and via computational experiments.