Task Value Calculus: Multi-Objective Trade off Analysis Using Multiple-Valued Decision Diagrams

Abstract

Most multiple-objective optimization algorithms utilize continuous input variables. Given that many decision variables in common use-cases are discrete rather than continuous, we develop a multiple-objective optimization framework over discrete variables known as task value calculus (TVC). The underlying mathematical models in TVC utilize a multiple-valued algebraic framework where both the objective functions and the system or process structure models are represented as multiple-valued functions. TVC allows for fast multiple-objective optimization through the use of the multiple-valued decision diagram (MDD) data structure. The algorithms and structures internal to TVC are described and experimental results are provided. TVC is implemented with a simple graphical user interface making it suitable for use by both laypersons and domain experts.