A chance constraint estimation approach to optimizing resource management under uncertainty

Abstract

Chance-constrained optimization is an important method for managing risk arising from random variations in natural resource systems, but the probabilistic formulations often pose mathematical programming problems that cannot be solved with exact methods. A heuristic estimation method for these problems is presented that combines a formulation for order statistic observations with the sample average approximation method as a substitute for chance constraints. The estimation method was tested on two problems, a small fire organization budgeting problem for which exact solutions are known and a much larger and more difficult habitat restoration problem for which exact solutions are unknown. The method performed well on both problems, quickly finding the correct solutions to the fire budgeting problem and repeatedly finding identical solutions to the habitat restoration problem.