Objective-CP is an optimisation system that views an optimisation program as the combination of a model, a search, and a solver. Models in Objective-CP follow the modelling style of constraint programming and are concretised into specific solvers. Search procedures are specified in terms of high-level nondeterministic constructs, search combinators, and node selection strategies. Objective-CP supports fully transparent parallelisation of multi-start and branch & bound algorithms. The implementation of Objective-CP is based on a sequence of model transformations, followed by a concretisation step. Moreover, Objective-CP features a constraint-programming solver following a micro-kernel architecture for ease of maintenance and extensibility. Experimental results show the practicability of the approach. We will develop plug-in modules extending the Objective-CP micro-kernel to incorporate state-of-art learning capabilities, and specific modelling and solving capabilities for important application areas such as scheduling and packing.
Research unit: Optimisation Platforms