Framework for Optimizing the Maintenance and Rehabilitation Schedules of Interdependent Infrastructure Systems

Majed Alinizzi, Purdue University

Abstract

In current practice, civil infrastructure systems are typically managed and operated without duly accounting for the interdependencies that exist among different types of systems. Managing infrastructure systems separately may result in practices that may be cost-effective locally (for the individual system only) but not globally (for two or more systems with shared locations or functionality). This dissertation develops an optimization framework for scheduling infrastructure repairs that considers such interdependencies. The framework determines the optimal performance thresholds for applying a treatment to a primary system that is subject to disruption induced by a neighboring (secondary) system. This disruption may be deterministic or probabilistic in terms of its severity and/or the time of its occurrence. The optimization framework considers both the incremental costs and benefits of treatments applied to the primary system. A number of techniques, including first-order derivative, multiobjective genetic algorithm (NSGA-II), and sample average approximation, are presented to solve the scheduling optimization problem. The developed framework is demonstrated using two case studies involving co-located pavement and underground utility assets. In the first case study, the primary system is the pavement to be maintained using a common rehabilitation treatment (i.e., a thin hot mix asphalt overlay) that is subject to utility cuts imposed by underground utilities. The aim is to schedule the pavement treatment such that the impacts of the utility cuts are minimized. In the second case study, a water pipe is the primary system that is located beneath a pavement, and replacement of the pipe is subject to permit fees imposed by the agency managing the overlying pavement. The goal is to schedule the pipe replacement such that the pavement cut permit fees are minimized. The results obtained from the case studies show that the proposed model can provide effective maintenance schedules for a primary system in the presence of disruptions due to a secondary system.

Degree

Ph.D.

Advisors

Samuel, Purdue University.

Subject Area

Civil engineering

COinS