A Highly Functional Decision Paradigm Based on Nonlinear Adaptive Genetic Algorithm
Abstract
In response to DOD Army solicitation A93-023, POC proposed a new type of processing module that is superior to any other type of computing method. The greatest advantage of POC's processing module is it combines the ability of a human to adapt to changing facts with the constant, uniform behavior, predictability, speed, large information storage capacity and non-emotional responses of machines. This proposed highly dimensional decision module is based on the genetic algorithm (GA). In Phase 1 of this program, POC successfully demonstrated a universal decision making method based on GAs and fuzzy rules. The algorithm of this decision making method can be used for multi-sensor feature space data, high dimensionality, and dynamically changing environments. In addition, the algorithm provides the capability to build a nonlinear adaptive system that incorporates many of the properties of human cognitive method The multidimensional (10 dimensions) function demonstration of the optimization algorithm in Phase 1 has fully proved the feasibility of POC's high dimensionality decision making approach. Based on these results, POC believes that the high dimensionality decision module can be fully implemented with existing optic or electronic components in Phase 2.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 20, 1994
- Accession Number
- ADA281457
Entities
People
- Dai H. Kim