Date of Graduation

5-2018

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level

Graduate

Department

Industrial Engineering

Advisor/Mentor

Pohl, Edward A.

Committee Member

Rossetti, Manuel D.

Second Committee Member

Rainwater, Chase E.

Third Committee Member

Scott, Marc A.

Keywords

Empty Container Management; Network Design; Transportation; Unmanned Aerial Vehicle

Abstract

In this dissertation, we address different transportation problems. The three main outcomes are: designing a battery swap station network, studying gaps in Empty Container Management literature, designing a model with similar characteristics to the vehicle routing problem as well as implementation scenario for it to be applied for a real-world case.

For the designed battery swap station, a model is developed for customer demand satisfaction that permits construction of different types of BSS in the planning network. Our solution methodology is a Tabu Search algorithm combined with a dynamic programming initialization. Numerous tests showed that the proposed TS approach provides improvement compared to CPLEX both in terms of time and solution quality.

The comprehensive literature review on Empty Container Management resulted in realizing intermodal environment as problem areas not being investigated much. Moreover, limited number of research has targeted large problem instances. The outcome of this investigation was to develop a demand forecasting model over the empty containers, a DSS for daily decision making, and a mathematical model to optimize the problem. Case studies with large instances of even more than 500 nodes as well as customization of the approach for a major transportation company in the US prove the applicability of the outcomes.

The final outcome of the presenting dissertation aims at implication of a novel process for TB diagnostic process. For this reason, a simulation model is developed that quantifies the impact of different scenarios which are feasible options to be executed. However, our investigation resulted that among different scenarios, TBOnDemand, in which patients request a vehicle to take their sample at their location, provides significant improvements. We developed a model with similar characteristics to vehicle routing problem with predetermined, but dynamic sample processing capacity allocated to each vehicle. Results of the solutions gained by CPLEX an offline assignment of vehicles to patients showed substantial reduction in serving time.

As stated above, the contribution of presenting dissertation includes but is not limited to introducing novel mathematical models, developing heuristic approaches, simulating uninvestigated problem area with largescale samples, constructing DSS tool with customized case studies.

Download

Share

COinS