GENET is a connectionist approach to constraint satisfaction. It is a
stochastic search method. It has been shown to be both efficient and effective
in binary constraint satisfaction problems, graph colouring and car sequencing
(which is a non-binary problem).
Guided Genetic Algorithm (GLS) is an extension of GLS. It is a hybrid
between GLS and GA. The aim is to improve the robustness of GLS and to extend
its domain of application. It has been shown efficient and effective in a number
of problems, including the Royal Road Function, the Processors configuration
problem, the frequency assignment problem and the general assignment problem.
A system for modelling was developed in the
CACP Project.
A model of a constraint satisfaction problem is defined by the variables,
domains and contraints selected. Given a model specificaiton, the choice of
model is crucial to the solving of a constraint satisfaction problem. This
project looks at heuristics to evaluate models and modify it in order to help
solving constraint satisfaction problem.
This project is built upon the philosophy of finding appropriate algorithms
for a given constraint satisfaction (as opposed to applying a "champion
algorithm" to all problems). Mechanisms are being developed to monitor the
performance of algorithms and adaptive strategies are being developed to
dynamically switches algorithms when the current algorithm is concluded failing.
This is a completed project which aims to bring constraint technology to
non-expert users. The software engineering and human-computer interaction
aspects of constraint satisfaction and constraint optimization will be
investigated. Practical aspects of the constraint technology will be advanced. A
computer-aided constraint-programming system will be designed and implemented.
This research concerns itself with the scheduling of Autonomous Guided Vehicles
(AGVs) in a port. Port components that are relevant to our problem include
berths, quay cranes, container storage areas, and a road network. Given a number
of AGVs and their availability, the task is to schedule the AGVs to meet the
transportation requirements. We have extended classical Network Simplex Method
for (a) efficiency, and (b) dynamic problems. We have also developed a heuristic
search algorithm for this problem.
Population-based algorithms using estimation of distribution, often called
estimation of distribution algorithms (EDAs), have been recognized as a major
paradigm in evolutionary computation. This project will work towards
establishing a sound theory for characterizing and explaining EDA-like
algorithms. We shall use continuous optimization problems and quadratic
assignment problems as our test problems.
This project is in collaboration with BT. Although it is motivated by BT's
workforce scheduling problem, the ideas developed in this project are general.
It involved scheduling engineers to jobs, satisfying a wide range of
constraints. This is a multi-objective optimization problem. Some of the
objectives are to minimize travelling distance and to maximize service quality
as defined by the company. Staff empowerment is also a major theme in this
project.