Genetic-algorithm programming of road maintenance and rehabilitation

Abstract

This paper describes the development of a computer model (known as PAVENET-R) based on genetic algorithms, an optimization tool capable of overcoming combinatorial explosion, to solve the pavement maintenance-rehabilitation trade-off problem at the network level. The formulation of the PAVENET-R model is described in detail. An integer coding scheme is selected for parameter representation in the model. Two genetic-algorithm operators, namely the crossover operator and the mutation operator, are used. A change table encodes constraints to the genetic-algorithm operations to ensure that only valid offspring are generated from a parent pool. Four numerical examples of road networks of 30 pavement segments, each with different relative costs of rehabilitation and maintenance activities, are analyzed to demonstrate the trade-off relationship between pavement rehabilitation and maintenance activities. The detailed maintenance and rehabilitation schedules of the solutions, and the convergence characteristics of each solution are presented.