Genetic Algorithms–Based Network Optimization System with Multiple Objectives

In any engineering decision-making process, it is a common problem to formulate a model with multiple objectives and to obtain proper answers. If the objectives are opposing, the problem then becomes finding the best possible design that identifies the best compromise between the opposing objectives. In a pavement management system of a state highway agency, engineers constantly struggle with performance requirements and limited budget. This situation results in an optimization problem with two conflicting objectives. The underlying algorithm of the classic network optimization system was coupled with genetic algorithm (GA) techniques to minimize budget and maximize performance of highway pavement networks in rehabilitation planning. GA is generally considered a powerful methodology in solving multiobjective problems. The data analysis demonstrated that the GA technique is applicable in producing workable solutions to a multiobjective pavement management problem.