My research focuses on utilizing Network Optimization, Integer Programming, Robust Optimization and Bilevel programming to solve transportation and logistics problems, particularly the hazardous material network design problem.
Hazardous Material Network Design
Accidents involving transporting hazardous materials (hazmat) could cause catastrophic damages and consequences in means of both human casualties and residential environments. In the year of 2014, there are 4,565 incidents during the transit of hazmat in the United States, resulting in over $70 million total cost, 56 injuries and 8 fatalities. It is important for authorities to regulate hazmat transportation in urban traffic networks.
In a vehicular network, a risk is associated with each road segment depending on factors such as hazmat type and population around the road. The government aims to minimize the risk in transporting hazmat while carriers are more concerned with cost. In reality, minimum cost and minimum risk routes are usually different. Due to regulation restrictions, the government is not able to enforce carriers to choose the minimum risk routes. Instead, the government regulates hazmat transportation on the network under its jurisdiction by closing certain road segments or enforcing tolls. My research focuses on the regulation method of closing road segments. Then we have a hazmat network design problem: the government selects the road segments that should be closed so as to minimize total risk, considering the behaviors that carriers will then choose the minimum cost routes on the resulting network.
In current literature regarding hazmat network design problems, very few works deal with uncertainty. However uncertainty could be pervasive and leads to an undesirable solution. We consider two kinds of uncertainties:
- Parameter measures (risk) uncertainty
- Uncertainty in the route choice of the hazmat carriers
Based on the above, we propose robust hazmat network design models. Another problem we consider in this context is how to incorporate the cost equity among carriers in designing the road network.
For hazmat incidents, a good source of data is from US department of transportatioin, pipeline and hazardous materials safety administration.