►►►►Typical Projects

Constraint Reasoning and Constraint Programming (Supported by the National Natural Science Foundation of China under grant Nos.60073039, 60273080) The Key Laboratory of Symbolic Computation & Knowledge Engineering, Ministry of Education, China

Product Configuration Project Based on Constraint Reasoning (Supported by the Natural Science Foundation of Jilin Province of China under grant No.20040526) The Key Laboratory of Symbolic Computation & Knowledge Engineering, Ministry of Education, China [configurator]

Contributions to the Project: Proposed a new approach (I defined a tool called Truth Graph) to solve logical constraints efficiently, which is the core problem faced by our research group. Found an effective way to convert the logical constraints to linear algebraic equations, which made the NP problem easy to solve, also with efficiency. [SUMMARY]

Algorithms and Applications of Constraint Programming on Airport Gate Assignment problem [REPORT]

We try to build an effective model for the airport authorities to estimate the efficiency of the current airport gate assignment. This work is related to Continental Airlines. 

Empirical Study on Functional Constraints: Algorithm and Implementation. [Report]

In this project, we handle randomly generated CSPs, including a proportion a functional constraints. We employ the new technique of variable elimination to solve functional constraints first, which dramatically speeds up the solving process, compared to the traditional constraint solvers.

Integration of Constraint Programming and Integer Programming 

We try to integrate the technique of Constraint Programming and the advantages of Integer Programming  to provide a new powerful tool that can be more widely applied into combinatorial optimization problems.

Branch-and-Cut Approaches for Nonlinear and Mixed-Integer Nonlinear Programming. National Science Foundation, CMMI-0620755

We use the piecewise linear functions to transform original MIP to the Nonlinear programs and further to improve the state-of-the-art MIP algorithm. [Research Summary]