Joint Optimization of Capital Intensity and Workforce Flexibility in Distribution Centers Under Demand Uncertainty

Abstract

Distribution centers increasingly compete on the combined strength of two operating levers: the capital intensity of the facility, expressed through conveyor, sortation, goods-to-person, automated storage and retrieval, and robotic systems, and the flexibility of the workforce that runs it, expressed through the mix of permanent cross-trained associates and flexible or contingent labor. These two levers are typically planned by separate functions on separate timelines, even though both exist to absorb the same underlying source of risk: uncertain order volume. This paper develops a two-stage stochastic optimization model that jointly determines the capital intensity of a distribution center and the size of its core workforce before demand is realized, with flexible labor deployed as a recourse action once demand becomes known. A closed-form, newsvendor-type critical fractile characterizes the optimal core workforce for any given capital intensity, and the resulting expected cost function is minimized over capital intensity using a nested search procedure. Closed-form expressions for the expected recourse cost are obtained under a lognormal demand assumption, allowing exact evaluation without simulation. A synthetic numerical study, calibrated to representative ranges for large scale, multi-channel distribution operations, shows that jointly optimizing capital intensity and workforce flexibility reduces total expected annual cost by approximately eight percent relative to a capital investment decision made without access to workforce flexibility, and by more than fifty percent relative to a workforce only strategy that forgoes automation investment. Optimal capital intensity is shown to increase with demand uncertainty and with the wage premium commanded by flexible labor, and to decrease with the cost of capital. The results offer a quantitative basis for coordinating distribution center engineering and workforce planning decisions and identify the conditions under which each lever substitutes for, rather than complements, the other.