Computerized System for Efficient Delivery of Infrastructure Maintenance/Repair Programs

Abstract

This paper discusses three common approaches used by owner organizations to deliver infrastructure maintenance/repair (M/R) programs that involve multiple distributed sites: using in-house resources, outsourcing to contractors, and a combination of both. A scheduling model is then introduced that uses genetic algorithms to suggest the optimum combination of in-house crews and outsourcing that meet execution constraints. The scheduling model considers all the variables that relate to in-house delivery, such as available crews, methods of construction, sequence of work sites, and the time and cost of moving resources among sites. In addition, it incorporates a mechanism to examine the feasibility of replacing in-house resources by outsourcing to delegate execution risks, at minimum additional cost. Details of model development are discussed and an example application is presented to demonstrate the planning and control features of the proposed model. The model presents practitioners and researchers with a practical tool for realistic what-if assessment and greater level of control over the execution of infrastructure M/R plans.