The goal of ECSPLAIN is to develop methods, techniques and generic software for modelling and solving rich models of complex, industrial constrained optimisation tasks. More specifically, the project focuses on the resolution of difficult problems often "over-constrained" involving multiple optimisation criteria and/or a wide variety of preference constraints. Results, in the form of more powerful Constraint Solver Intelligent Tools, and a methodology for mixed-initiative resolution of complex, constrained optimisation tasks, will be widely published, exploited and commercialised by the ECSPLAIN consortium.
1) To keep in Europe the world leadership in constraint technology. ECSPLAIN will develop methods, techniques and generic software for addressing over-constrained problems and complex industrial problems involving multiple optimisation criteria and/or a wide variety of preference constraints, ECSPLAIN will enable complex industrial problems to be described explicitly, allowing the modelling of high-level requirements, multiple optimisation criteria and preferences to be captured.
2) To improve the efficiency and quality in several areas of industrial activity. The developed methods and software will be applicable to a wide variety of optimisation problems with preferences, including construction planning and forestry management, time-tabling, resource allocation and scheduling.
3) To promote the project results and to strengthen IT businesses co-operation with Mediterranean and Middle East countries.
ECSPLAIN proposes to start from analysis of two industrial situations that are representative of complex applications requiring preference manipulation, and to analyse and generalise their overall needs, their user interfacing and interaction requirements, and their measures of success. In-house theoretical and practical experience with application of the constraint approach to problem solving will provide the foundations for methodological guidelines to constraint satisfaction with preferences, augmented by input from the users as seen above. In-house and academic know-how on constraint satisfaction with preferences will be employed in active research on existing algorithms and their combinations, and will pave the way for goal-driven research into new algorithmic approaches. Application of the first basic algorithmic building-blocks and rough methodological precepts by highly motivated in-house user partners will permit first-hand testing and feedback. Within the ECSPLAIN consortium, the presence of a commercial constraint solver will provide immediate leverage for exploitation of results. ECSPLAIN will thus build on the state-of-the-art, employing partner expertise in industrial application of constraint programming, in algorithms and theory, and in methodology, plus two representative industrial application situations, to address a difficult aspect of the constraint approach (preferences), with results articulated into rapidly marketable form.
The following results are expected:
1) new algorithms integrated with state-of-the-art commercialised optimisation products using the CSP approach,
2) a methodology for mixed-initiative resolution of complex, constrained optimisation tasks with an accompanying practical guide to modelling & solving over-constrained problems,
3) a set of public benchmarks,
4) running prototypes for two industrial applications,
5) publications and participation / organisation of mixed industrial / academic workshops.
Funding SchemeCSC - Cost-sharing contracts