Esprit Project 22165 - CHIC-2
Creating Hybrid Solutions for Industry and Commerce
Technical - constraint logic programming , hybrid algorithms , mathematical programming , stochastic repair .
Sector - car industry , manufacturing , finance and banking , energy , steel industry .
High quality planning, scheduling and resource allocation lie at the core of many business operations. Their importance is growing with the increase in complexity of modern business, and the need for efficient response to international competition. The computational effort required to solve problems in planning and scheduling typically grows very rapidly with the size of the problem, currently making the solution of large problems impractical. CHIC-2 aims to extend the range of solvable problems and reduce the time and cost of producing good solutions.
These considerations set the agenda of CHIC-2:
CHIC-2 will tackle four applications, each supplied by a major end user in the project. The applications, which are of strategic business importance, are production planning (Renault), equipment leasing (Bouygues), electricity trading (Eastern Electric) and plant scheduling (Hellenic Steel) and illustrate the blending of combinatorial optimisation and domain-dependent constraints.
The project aims to deliver solutions which improve significantly on current solutions with respect to cost, quality and integration as well as to use the results to develop a methodology which will support the development of a wide range of new applications.
The principal objective is to design and deliver a range of hybrid algorithms for complex optimisation problems. These will be based on the integration of three different approaches, viz. constraint logic programming, mathematical programming and stochastic repair techniques. These approaches have sometimes been presented as competitors, but more recently, they have come to be seen as complementary. CHIC-2 is dedicated to exploiting their complementary strengths.
The need for hybrid algorithms to address optimisation problems is due to the fact that different techniques are suited to solving different parts or aspects of the problems. This can be put another way. Many optimisation problems are hybrid problems, that embrace several different requirements, e.g. matching, routing and scheduling (real-life dispatching problems). In such a case there is a clear need to mix the techniques, that is, to build hybrid algorithms.
To facilitate the efficient design and development of hybrid algorithms the Project will seek to specify (1) a uniform modelling of complex optimisation problems and (2) a methodology for mapping the resulting models onto appropriate hybrid algorithms.
This strategy promises to yield much improved solutions to the optimisation problems of the user organisations. This will be realised through the effective integration of a wide variety of different but related aspects of the problems. Another envisaged benefit of the methodology is that it should reduce both the time and cost of handling new problems.
CHIC-2 will deliver a new programming environment ECLiPSe.2 that is designed to support the efficient implementation and execution of the hybrid algorithms. The environment is to be a successor to the ECLiPSe technology, developed on the Esprit CHIC project (P5291).
The two primary aims in developing ECLiPSe.2 are (1) to provide an efficient, flexible programming environment for prototyping and developing applications software addressing large scale optimisation problems; and (2) to provide a reliable, robust platform for delivering solutions and embedding the software in existing corporate environments.
Mr. Alan Ainsworth
Imperial College of Science, Technology and
Medicine - IC Parc
London SW7 2AZ
E-mail: (E-mail removed)