Abstract
The increasing competition faced by logistics carriers requires them to ship at lower cost and higher efficiency. In reality, however, many trucks are running empty or with a partial load. Bridging such residual capacity with real time transportation demand enhances the efficiency of the carriers. We therefore introduce the Position-Aware-Market (PAM), where transportation requests are traded in real time to utilize transportation capacities optimally. In this paper we mainly focus on the decision support system for the truck driver, which solves a profit- maximizing Pickup and Delivery Problem with Time Windows (PM-PDPTW). We propose a novel Recursive Branch-and-Bound algorithm that solves the problem optimally, and apply it to a Tabu-Search heuristic for larger problem instances. Simulations show that problems with up to 50 requests can be solved optimally within seconds. Larger problems with 200 requests can be solved approximately by Tabu-Search in seconds, retaining 60% of the optimal profit.
Recommended Citation
Qiu, Xiaoqiu; Neumann, Dirk; and Goetzinger, Markus, "The Position-Aware-Market: Optimizing Freight Delivery for Less-Than-Truckload Transportation" (2012). AMCIS 2012 Proceedings. 5.
https://aisel.aisnet.org/amcis2012/proceedings/GreenIS/5
The Position-Aware-Market: Optimizing Freight Delivery for Less-Than-Truckload Transportation
The increasing competition faced by logistics carriers requires them to ship at lower cost and higher efficiency. In reality, however, many trucks are running empty or with a partial load. Bridging such residual capacity with real time transportation demand enhances the efficiency of the carriers. We therefore introduce the Position-Aware-Market (PAM), where transportation requests are traded in real time to utilize transportation capacities optimally. In this paper we mainly focus on the decision support system for the truck driver, which solves a profit- maximizing Pickup and Delivery Problem with Time Windows (PM-PDPTW). We propose a novel Recursive Branch-and-Bound algorithm that solves the problem optimally, and apply it to a Tabu-Search heuristic for larger problem instances. Simulations show that problems with up to 50 requests can be solved optimally within seconds. Larger problems with 200 requests can be solved approximately by Tabu-Search in seconds, retaining 60% of the optimal profit.