Research on NOLH Experimental Designs with Discrete Factors for Multiple-Agent Simulation Based on SEAS

Abstract

Highly efficient and effective experimental methods designed for military, aerospace engineering, bioengineering and medicine purposes have been applied in the fields for a long time. Based on the foundation of uniform design and orthogonal Latin hypercube design methods, this paper is trying to make some improvements and introduce a new method to construct experimental design suitable for either discrete or categorical variables. The new method is testified and proved to be fast, simplified and efficient for experiment including numerous factors, which have complex interrelated relationship within each other.