Work

Computing explanations for interactive constraint-based systems

Year: 2011
Type: dissertation
Abstract: Constraint programming has emerged as a successful paradigm for modelling\ncombinatorial problems arising from practical situations. In many of those situations,\nwe are not provided with an immutable set of constraints. Instead, a user\nwill modify his requirements, in an interactive fashion, until he is satisfied with\na solution. Examples of such applications include, amongst others, model-based\ndiagnosis, expert systems, product configurators.\nThe system he interacts with must be able to assist him by showing the consequences\nof his requirements. Explanations are the ideal tool for providing this\nassistance. However, existing notions of explanations fail to provide sufficient information.\nWe define new forms of explanations that aim to be more informative.\nEven if explanation generation is a very hard task, in the applications we consider,\nwe must manage to provide a satisfactory level of interactivity and, therefore, we\ncannot afford long computational times.\nWe introduce the concept of representative sets of relaxations, a compact set of\nrelaxations that shows the user at least one way to satisfy each of his requirements\nand at least one way to relax them, and present an algorithm that efficiently computes\nsuch sets. We introduce the concept of most soluble relaxations, maximising\nthe number of products they allow. We present algorithms to compute such relaxations\nin times compatible with interactivity, achieving this by indifferently making\nuse of different types of compiled representations. We propose to generalise\nthe concept of prime implicates to constraint problems with the concept of domain\nconsequences, and suggest to generate them as a compilation strategy. This sets a\nnew approach in compilation, and allows to address explanation-related queries in\nan efficient way. We define ordered automata to compactly represent large sets of\ndomain consequences, in an orthogonal way from existing compilation techniques\nthat represent large sets of solutions.
Language: English
Open Access status: green
OpenAlex ID: W2183657870
Cites: 131
Cited by: 1
Related to: 10