Genetic-Algorithm-Based Normal Boundary Intersection (GANBI) Method; An Efficient Approach to Pareto Multiobjective Optimization for Engineering Design

摘要

A new method for developing tradeoffs in the engineering of complex systems is described. The Genetic-Algorithm-Based Normal Boundary Intersection (GANBI) method serves as a preprocessor for conventional genetic-algorithm-based Pareto optimization solvers. The algorithm is based on applying the normal boundary intersection approach of Pareto optimization to genetic solvers. The approach is shown to provide rapid convergence and to provide a better estimate of the Pareto set than existing state-of-the-art methods. A description of Pareto optimization methods for engineering design is included to put the new method in the context of existing solution approaches. The algorithm for the GANBI method is derived and detailed in the report, and numerical examples showing its efficiency in solving an academic problem are presented. The report concludes with an example of how the GANBI method has been used to make tradeoff decisions in the design of large-scale distributed undersea sensor networks.