A Stochastic Bicriteria Procedure for Creating System Options: A Recent Study

Stochastic systems are sometimes swamped by incompatible performance needs and inconsistent performance specifications, which can be difficult to discover when supporting decision models must be built. As a result, it is frequently desirable to provide a set of disparate solutions that allow for different approaches to the problem. These options should meet the system's performance requirements while remaining as distinct as possible in their decision spaces. This work presents a stochastic bicriteria approach for generating sets of maximally diverse options. In a single computer run of the technique, this stochastic algorithmic approach is both computationally efficient and concurrently creates the pre-specified number of maximally distinct solution alternatives. The efficacy of this stochastic MGA method is demonstrated on a “real world” water resources case study.