\(P_a\)-quacq: Algorithm for Constraint Acquisition System

Constraint acquisition system assists a non-expert user in modeling her problem as a constraint network. T-quacq is a recent learner system that acquires constraint network by generating queries in an amount of time not exceeding a waiting time upper bound. The hindrance is the risk of reaching a premature convergence state. In this paper we present \(P_a\)-quacq, a new algorithm based on T-quacq. \(P_a\)-quacq is able to made T-quacq more efficient in terms of time and convergence. Finally, we give an experimental evaluation of our algorithm on some benchmarks.

[1]  Eugene C. Freuder,et al.  Suggestion Strategies for Constraint-Based Matchmaker Agents , 2002, Int. J. Artif. Intell. Tools.

[2]  Barry O'Sullivan,et al.  Query-Driven Constraint Acquisition , 2007, IJCAI.

[3]  Eugene C. Freuder.  Constraint Acquisition , 2002, AISC.

[4]  Carine Lallemand,et al.  Enhancing User eXperience during waiting time in HCI: contributions of cognitive psychology , 2012, DIS '12.

[5]  Arnaud Lallouet,et al.  On Learning Constraint Problems , 2010, 2010 22nd IEEE International Conference on Tools with Artificial Intelligence.

[6]  Toby Walsh,et al.  New Approaches to Constraint Acquisition , 2016, Data Mining and Constraint Programming.

[7]  Toby Walsh,et al.  Constraint Acquisition via Partial Queries , 2013, IJCAI.

[8]  Nadjib Lazaar,et al.  Multiple Constraint Acquisition , 2015, CPCR+ITWP@IJCAI.

[9]  Christian Bessiere,et al.  Time-Bounded Query Generator for Constraint Acquisition , 2018, CPAIOR.

[10]  Gerhard Friedrich,et al.  Argumentation Based Constraint Acquisition , 2009, 2009 Ninth IEEE International Conference on Data Mining.

[11]  Helmut Simonis,et al.  A Model Seeker: Extracting Global Constraint Models from Positive Examples , 2012, CP.