Towards Portfolios of Streamlined Constraint Models: A Case Study with the Balanced Academic Curriculum Problem

Augmenting a base constraint model with additional constraints can strengthen the inferences made by a solver and therefore reduce search effort. We focus on the automatic addition of streamliner constraints, derived from the types present in an abstract Essence specification of a problem class of interest, which trade completeness for potentially very significant reduction in search. The refinement of streamlined Essence specifications into constraint models suitable for input to constraint solvers gives rise to a large number of modelling choices in addition to those required for the base Essence specification. Previous automated streamlining approaches have been limited in evaluating only a single default model for each streamlined specification. In this paper we explore the effect of model selection in the context of streamlined specifications. We propose a new best-first search method that generates a portfolio of Pareto Optimal streamliner-model combinations by evaluating for each streamliner a portfolio of models to search and explore the variability in performance and find the optimal model. Various forms of racing are utilised to constrain the computational cost of training.

[1]  Jimmy Ho-Man Lee,et al.  Increasing Constraint Propagation by Redundant Modeling: an Experience Report , 1999, Constraints.

[2]  Leslie Pérez Cáceres,et al.  The irace package: Iterated racing for automatic algorithm configuration , 2016 .

[3]  John R. Rice,et al.  The Algorithm Selection Problem , 1976, Adv. Comput..

[4]  Brahim Hnich,et al.  Extensible Automated Constraint Modelling , 2011, AAAI.

[5]  Bart Selman,et al.  On the Erd˝ os Discrepancy Problem* , 2014 .

[6]  Michèle Sebag,et al.  Hypervolume indicator and dominance reward based multi-objective Monte-Carlo Tree Search , 2013, Machine Learning.

[7]  Ian Miguel,et al.  Instance Generation via Generator Instances , 2019, CP.

[8]  Kevin Leyton-Brown,et al.  Hydra: Automatically Configuring Algorithms for Portfolio-Based Selection , 2010, AAAI.

[9]  Marius Thomas Lindauer,et al.  AutoFolio: An Automatically Configured Algorithm Selector , 2015, J. Artif. Intell. Res..

[10]  Ian Miguel,et al.  Automatic Streamlining for Constrained Optimisation , 2019, CP.

[11]  Özgür Akgün.  Extensible automated constraint modelling via refinement of abstract problem specifications , 2017 .

[12]  Ian Miguel,et al.  Automatically Generating Streamlined Constraint Models with Essence and Conjure , 2015, CP.

[13]  Simon M. Lucas,et al.  A Survey of Monte Carlo Tree Search Methods , 2012, IEEE Transactions on Computational Intelligence and AI in Games.

[14]  Kevin Leyton-Brown,et al.  Sequential Model-Based Optimization for General Algorithm Configuration , 2011, LION.

[15]  Warwick Harvey,et al.  Essence: A constraint language for specifying combinatorial problems , 2007, Constraints.

[16]  Bart Selman,et al.  On the Erdős Discrepancy Problem , 2014, CP.

[17]  Holger H. Hoos,et al.  Automated Algorithm Configuration and Parameter Tuning , 2012, Autonomous Search.

[18]  Toby Walsh,et al.  CGRASS: A System for Transforming Constraint Satisfaction Problems , 2002, International Workshop on Constraint Solving and Constraint Logic Programming.

[19]  Stephen P. Schwartz,et al.  The Essence of Essence , 2009 .

[20]  Ian Miguel,et al.  Automatic Generation and Selection of Streamlined Constraint Models via Monte Carlo Search on a Model Lattice , 2018, CP.

[21]  Simon Colton,et al.  Constraint Generation via Automated Theory Formation , 2001, CP.

[22]  Greg Hamerly,et al.  Learning the k in k-means , 2003, NIPS.

[23]  Peter Nightingale,et al.  Essence' Description , 2016, ArXiv.

[24]  Ian Miguel,et al.  Automatically Improving SAT Encoding of Constraint Problems Through Common Subexpression Elimination in Savile Row , 2015, CP.

[25]  Maurizio Gabbrielli,et al.  An enhanced features extractor for a portfolio of constraint solvers , 2013, SAC.

[26]  Meinolf Sellmann,et al.  Streamlined Constraint Reasoning , 2004, CP.

[27]  Peter J. Stuckey,et al.  MiniZinc: Towards a Standard CP Modelling Language , 2007, CP.

[28]  Heike Trautmann,et al.  Automated Algorithm Selection: Survey and Perspectives , 2018, Evolutionary Computation.

[29]  Bilal Syed Hussain,et al.  Automated Symmetry Breaking and Model Selection in Conjure , 2013, CP.

[30]  Peter J. Stuckey,et al.  Exploiting subproblem dominance in constraint programming , 2011, Constraints.

[31]  Ian P. Gent,et al.  Automatically improving constraint models in Savile Row , 2017, Artif. Intell..

[32]  Lars Kottho,et al.  Algorithm Selection for Combinatorial Search Problems: A survey , 2012 .

[33]  Simon Colton,et al.  Automatic Generation of Implied Constraints , 2006, ECAI.

[34]  Ronan Le Bras,et al.  Double-Wheel Graphs Are Graceful , 2013, IJCAI.

[35]  Ian P. Gent,et al.  Breaking Conditional Symmetry in Automated Constraint Modelling with CONJURE , 2014, ECAI.

[36]  Michal Kouril,et al.  Resolution Tunnels for Improved SAT Solver Performance , 2005, SAT.

[37]  Thomas Stützle,et al.  A Racing Algorithm for Configuring Metaheuristics , 2002, GECCO.

[38]  Leslie Pérez Cáceres,et al.  An Experimental Study of Adaptive Capping in irace , 2017, LION.

[39]  Lars Kotthoff,et al.  Algorithm Selection for Combinatorial Search Problems: A Survey , 2012, AI Mag..