TY - BOOK
T1 - Particle swarm optimisation of pavement maintenance and rehabilitation
AU - Wang, Yunzhong
PY - 2013
Y1 - 2013
N2 - Pavement project selection, prioritisation and scheduling need to be based on a systematic evaluation approach which takes into account strategic considerations as well as network level benefit justification. Since the benefits at the network level from various pavement maintenance and rehabilitation projects (pavement projects in short) are not additive, it is essential to consider project interdependence when planning such projects. Project interdependence can be either indirect, such as traffic interaction between projects or benefit interaction over the network, or direct, such as project alignment and coordination. Based on the results of this study, it is evident that both types of project interdependence have significant impacts on the quality of the optimised plan. The development of a multi-year multi-objective pavement planning optimisation framework in this study makes it possible for plans to take into account both direct and indirect project interdependence and also other objectives of interest to the policy maker or adviser. The framework systematically formalises the decision making process for pavement asset management in general and the pavement project planning problem in particular. It consists of a series of models including a Vector Evaluated Particle Swarm Optimisation (VEPSO) base model, a project initialisation and evaluation model, an HDM_III /4 based pavement deterioration model and a traffic updating model. An illustrative example based on the framework is presented. This takes the form of a proof of concept prototype. The prototype is a multi-objective multi-year pavement maintenance and rehabilitation project planning system, called MMPPS. For the purpose of comparison, the development also includes a single objective multiyear pavement maintenance and rehabilitation project planning prototype system, called SMPPS to reflect current practice in planning pavement projects. The only project interdependence taken into account in SMPPS optimisation is indirect interdependence. Both systems are applied to optimise a ten-year plan for maintenance and rehabilitation of the road network in the Great Southern Region of Western Australia. Based on the case study, the observation can be made that the optimisation which takes account of both direct and indirect project interdependence can significantly improve the quality of the suggested plan.
AB - Pavement project selection, prioritisation and scheduling need to be based on a systematic evaluation approach which takes into account strategic considerations as well as network level benefit justification. Since the benefits at the network level from various pavement maintenance and rehabilitation projects (pavement projects in short) are not additive, it is essential to consider project interdependence when planning such projects. Project interdependence can be either indirect, such as traffic interaction between projects or benefit interaction over the network, or direct, such as project alignment and coordination. Based on the results of this study, it is evident that both types of project interdependence have significant impacts on the quality of the optimised plan. The development of a multi-year multi-objective pavement planning optimisation framework in this study makes it possible for plans to take into account both direct and indirect project interdependence and also other objectives of interest to the policy maker or adviser. The framework systematically formalises the decision making process for pavement asset management in general and the pavement project planning problem in particular. It consists of a series of models including a Vector Evaluated Particle Swarm Optimisation (VEPSO) base model, a project initialisation and evaluation model, an HDM_III /4 based pavement deterioration model and a traffic updating model. An illustrative example based on the framework is presented. This takes the form of a proof of concept prototype. The prototype is a multi-objective multi-year pavement maintenance and rehabilitation project planning system, called MMPPS. For the purpose of comparison, the development also includes a single objective multiyear pavement maintenance and rehabilitation project planning prototype system, called SMPPS to reflect current practice in planning pavement projects. The only project interdependence taken into account in SMPPS optimisation is indirect interdependence. Both systems are applied to optimise a ten-year plan for maintenance and rehabilitation of the road network in the Great Southern Region of Western Australia. Based on the case study, the observation can be made that the optimisation which takes account of both direct and indirect project interdependence can significantly improve the quality of the suggested plan.
KW - Multiobjective optimization
KW - Vector evaluated particle swarm optimisation (VEPSO)
KW - Assert management
KW - Transport economics and planning
KW - Project direct and indirect interdependence
KW - Pavement project selection, prioritization and scheduling
KW - Information technology and systems
KW - Pareto-VEPSO model
M3 - Doctoral Thesis
ER -