The Heterogeneous Fleet Risk-Constrained Vehicle Routing Problem in Humanitarian Logistics

While distributing essential supplies in volatile environments, humanitarian transport is often exposed to threats such as attacks. To mitigate the negative consequences of attacks, we introduce the heterogeneous fleet risk-constrained vehicle routing problem (HFRCVRP), in which we aim to minimize transportation costs and the expected loss of getting robbed. An Adaptive Large Neighborhood Search (ALNS) heuristic is presented to solve the problem. The trade-off between transportation costs and expected loss of attacks is analyzed with a real-world case in South Sudan. Results show that the trade-off is especially relevant in the heterogeneous variant, in which Unmanned Aerial Vehicles (UAVs) can effectively mitigate risks of truck transport, providing a 7.1% improvement of the objective value compared to the same instance with only trucks. Risks can be completely eliminated by increasing transportation costs by a factor of five. Additionally, the risk variant decreases the objective value by 14.8% compared to considering only transportation costs and ignoring risks.