Evaluation of Methods for Generating Alternatives to Regional Wastewater Treatment Systems

Optimization models for complex regional planning problems can be used to generate planning alternatives that are good with respect to the objectives included in the models, but significantly different from each other with respect to the decisions specified, in order to provide more insight and understanding for the analyst or decisionmaker.

The objectives of the study are (1) to develop a methodology for the evaluation of a modeling-to-generate-alternatives (MGA) method, and (2) to test systematically the MGA methods for generating planning alternatives that are good and different. A regional wastewater treatment system problem, which is formulated as a mixed integer programming model, is used as an example. Solutions that are feasible and meet the specified cost target among all possible combinations of treatment plant locations are identified as the base population. Cluster analysis is then performed to allocate these solutions into a set of mutually exclusive groups. Each of the MGA methods is used to generate several sets of solutions and is evaluated on the basis of the percentage coverage of the groups of the base population. This criterion provides a further assessment of the use of these MGA methods.