Assessing the Impact of Driver Overtime in the Distribution Network of a Flower Retail Chain

Overview

The study quantifies effects of permitting driver overtime within a capacitated vehicle routing context informed by a flower retail chain's daily operations. The problem includes deliveries, split pickups, a heterogeneous fleet, and a heterogeneous driver workforce. The objective integrates route driving costs during ordinary hours, explicit overtime costs, and additional handling time for pickups; social constraints capture driver time workloads. Comparative experiments assess model-derived solutions against manually produced schedules and a commercial routing package, and contrast cost minimization with distance minimization formulations.

Methods and approach

A route-based mixed integer linear programming formulation was developed to represent routing choices, vehicle capacity, split pickups, per-driver time budgets, and explicit overtime allowance and cost parameters. The objective function comprises ordinary-hour driving costs, overtime wage costs, and pickup handling time penalties. Social constraints enforce driver workload limits and feasible overtime attribution. Computational experiments use real operational data from the retailer to evaluate alternative maximum overtime allowances and to compare outcomes to manually generated routings, a commercial solver baseline, and an alternative formulation where distance is minimized instead of cost.

Results

Model solutions yielded substantive cost reductions relative to manually produced plans (17.4%–36.4%) and relative to the commercial software baseline (9.7%–25.5%), with variation driven by overtime allowance levels and instance characteristics. Allowing overtime systematically improved cost performance, particularly for serving customers located at larger travel distances from the depot; overtime enabled consolidation and elimination of otherwise costly long-route repetitions. Replacing the cost objective with distance minimization produced routes that diverged notably from cost-optimal routes, indicating that distance-minimization can be a poor proxy when overtime and time-dependent labor costs are material.

Implications

Incorporating explicit overtime costs and driver workload constraints in routing formulations materially affects operational recommendations and realized cost outcomes; models that omit overtime or treat distance as the surrogate objective risk underestimating labor-driven costs. Optimal use of limited overtime can be concentrated on long-distance service nodes to capture the greatest marginal cost benefit, implying scheduling policies that prioritize overtime allocation by travel-time exposure. For planning and procurement of routing solutions, evaluation criteria should include labor cost structure and social constraints rather than relying solely on distance or classical distance-based heuristics.