Optimization Methodology for Multi-Criteria Comparison Using Response Surface Methods and Mathematical Programming

摘要

The maxima/minima of one response surface (measure of effectiveness, MOE) is calculated when the domain is constrained by consecutive levels (aspirations) of a second response surface. Thus, the optimal trade-off between the two responses (MOE's) is displayed over an entire operating domain, providing new and unique capabilities to evaluate options. The study applies this development to a strategic force structure modeled in a linear programming model. The methodology determines the optimal force structure to maximize one MOE given the aspiration levels of the other MOE. Therefore, the decision maker has the optimal force structures to achieve a range of demands/aspirations and also maximize the second goal. The methodology has applications to multicriteria optimization when relatively large numbers of domain control variable combinations are involved. This technique may be a tool to design the optimal weapon system characteristics or choose the optimal mix of weapon types to fulfill dual mission requirements. This work is an extension of previous work which used response surfaces, experimental design, and mathematical programming to conduct multi-dimensional, sensitivity analysis and weapon system comparisons. (Author)