Equity and network-level maintenance scheduling

Determining maintenance policies based purely on cost-minimization principles may result in an inequitable distribution of funding and benefits—in practice, such models often recommend allocating much more funding to urban areas than rural, which is arguably unfair and politically infeasible. This paper considers how preference for equitable distribution of benefits can be incorporated into network-level maintenance optimization problems. In particular, several seemingly natural choices for capturing inequity lead to nonconvex formulations. Accordingly, this paper investigates a class of functions for measuring inequity and identifies specifications which preserve convexity and lend themselves to easy optimization problems. These alternative specifications are compared on a fictitious network. Within this data set, considerable gains in equity are possible with only slight increases in total user cost.