| dc.description.abstract | Unexpected disruptive events have significant impacts on Supply Chain Networks (SCN), their capacity, and overall performance. Thus, it is critical to developing efficient techniques to improve their robustness and resilience while guaranteeing other constraints are met. This dissertation proposes four mathematical models that aim to improve the supply chain transportation network performance (SCTN). Chapter 2 proposes the first model, a bi-objective mixed integer programming (MIP) model that minimizes the impacts of disruptive events on the SCN while procuring a fairness-based distribution of commodities. To illustrate the proposed bi-objective MIP model, we present a realistic case study based on the transportation of commodities in Colombia. The case study considers a multi-commodity framework with 258 products and raw materials encompassing the country's six main economic sectors. The results show that seeking fairness within cost-effective distribution strategies reduces disruptions and helps maintain the required service levels at more demand points than when following distribution strategies that only focus on cost reduction.
Due to the high dependency on infrastructure networks, having an effective restoration plan to restore the network performance is essential. Chapter 3 proposes the second model, a multi-criteria nonlinear mixed integer programming (NMIP) model, to deal with disruptions in transportation networks. The model examines two restoration approaches to recover the transportation network performance, thus dependent networks, supply chain network (SCN), resilience improves. The model focuses on reducing the travel time of delivering the required demand in order to maintain a higher satisfaction rate. The model is designed to restore the transportation network during unpredicted events based not only on the cost-effectiveness approach but also on fairness. The results show that the fair distribution helps to maintain and recover the satisfaction rate faster when comparing to the results of the restoration plan based on the cost-effectiveness approach only.
The SCN is highly interconnected within its network; thus, for any disruption in any part of the SCN, the economic impacts might be catastrophic on the SCN overall performance. Consequently, in Chapter 4, we propose the third model, a bi-objective mixed integer programming (MIP) model that aims to minimize the impacts of disruptive events and restore the SCN performance, considering the economic interdependencies among various SCNs. The results show that both restoration methods helped to restore and recover the disrupted components and economic performances of the SCN. In addition, however, the fairness-based distribution method helped to keep most of the demand nodes satisfied by maintaining the required Service Level (SL) for the majority during the restoration time until the system had been fully restored.
As a result of the strong reliance on infrastructure networks, having efficient mitigation plans to protect the road transportation networks and restore them during the post-disruption stage is critical; thus, the network functionality is recovered. In Chapter 4, the fourth model, two bi-objective models, are proposed. The first model is a bi-objective, robust, fair mitigation MIP model that aims to allocate the mitigation budget fairly to generate a robust solution against the worst-case scenario that might impact the SCN performance. The other model is a bi-objective, fairness-based distribution MIP model that aims to restore the disrupted components in the road network and restore SCN performance to normal status. The results show that mitigating the road network before disruptions occur and implementing a fairness-based distribution during the post-disruption help SCNs recover quickly and maintain a higher service level at most demand nodes for each commodity. | en_US |
