A two-stage heuristic approach for fleet management optimization under time-varying demand

Abstract

An efficient two-stage heuristic approach is developed for solving the fleet management problem under time-varying demand. Stage 1 of the approach optimizes the vehicles’ utilization schedule. Continuous-time approximation is employed to yield a set of near-optimality conditions that can greatly reduce the solution space of this stage. Stage 2 then optimizes the vehicle purchase and retirement schedules. Numerical experiments showed that our approach outperformed a number of previous methods and commercial solvers by large margins in terms of solution quality, computational efficiency, or both.