Air Taxi Problem with Plane Changes

We developed a real time flight scheduling algorithm for small planes with five seats. Evaluated the extra benefit obtained by allowing plane changes. An intelligent local search heuristic and a mathematical formulation were developed for the problem. Computational experiments showed that around 8% of total flying cost was saved. The algorithm is implemented in C++ and I was involved in the project for 15 months. Project is supported by The Logistics Institute and Day-Jet Inc. Supervisors: Mokhtar Bazaraa, Martin Savelsbergh and George L. Nemhauser.

Real-Time Decision Support Infrastructure

We developed an engine library of a DSI (Decision Support Infrastructure) for transportation and logistics companies, which include different efficient local search algorithms, network simplex algorithm, Danzig-Wolfe decomposition, Lagrange relaxation algorithm and so on. The system is implemented in JAVA and the software is used by Jardin Logistics and other logistics companies in Hong Kong. I was involved in the project for two years and in charge of the intelligent part of the software. Supervisor: Raymond K. Cheung. In the following, we show several figures captured from our software.

The following figure shows the input interface of our software.

The following figure shows the database of our software.

The following two figures show the truck route demonstration of our software.

The following figure shows the route demonstration of our software for shipping containers.