Enhancement of Set-Based Design Practices Via Introduction of Uncertainty Through Use of Interval Type-2 Modeling and General Type-2 Fuzzy Logic Agent Based Methods.

Abstract

The goal of this research was to discern the effects of introducing uncertainty representation into a set-based design process with applications in ship design. The hypothesis was that introduction of design uncertainty would enhance the facilitation of set-based design practices. A presentation of three fuzzy logic agent based methods for facilitation of set-based ship design practices is offered. The first method utilized a type-1 fuzzy logic system to facilitate set-based design practices and possessed no uncertainty modeling. The next two methods included representation of design uncertainty in the set-based design space. Of these two methods, one utilized a novel approach that harnessed techniques of randomization to model an interval type-2 fuzzy logic system, the other method made use of general type-2 fuzzy logic methods that were well-known, but still relatively under-utilized in academics and industry when compared to type-1 fuzzy logic systems. Comparisons of the newly developed fuzzy logic systems with each other, and the type-1 agent based fuzzy logic system provided the basis for conclusions as to the effects of introducing uncertainty modeling into a set-based design process. The results of this experimental research have shown that the inclusion of uncertainty modeling in the set-based design process for the negotiation of design variables enhances the overall set-based design progression, especially when working with highly constrained designs. In the case of a highly constrained design, the type-1 fuzzy logic system was unable to promote set-convergence within the allotted experimental time without repeated design failures, while the use of uncertainty modeling allowed the interval type-2 modeling and general type-2 fuzzy logic systems to achieve feasible set-based design convergence. When performing a simplistic, loosely constrained design, all three fuzzy logic systems were capable of facilitating the principle practices of set-based design within the feasible solution space; specifically, the set-based practices of delaying design decisions and gradual reduction of the feasible solution space. This research has led to the enhancement of the set-based design process by providing capabilities to now represent uncertainty in the set-based design space though the use of either the newly developed interval type-2 or general type-2 fuzzy logic systems.