On Computing Budget Allocation for Finding the Simplest Good Enough Design

Abstract Many systems in nowadays follow not only physical laws but also man-made rules. These systems are known as discrete-event dynamic systems (DEDS's). Due to various advantages in practice, designs (or solution candidates) with low descriptive complexities (called simple designs) are usually preferred than complex ones when their performances are close. However, the descriptive complexity is usually nonlinear and takes discrete value, which makes traditional methods such as linear programming and gradient-based local search not applicable. Existing methods for simulation-based optimization (SBO) barely explore the preference on descriptive complexity and thus do not solve the problem efficiently. In this paper, the problem of computing budget allocation for finding the simplest good enough design is considered. Equal allocation (EA), optimal computing budget allocation (OCBA), and OCBA-m are compared over multiple numerical examples. Surprisingly, EA has good performance in general, though not necessarily optimal. All three methods are then applied to solve a node activation policy optimization problem in a wireless sensor network. EA again has good performance. We hope this work can shed some insight on how to find simple and good designs in general.