@inbook{69383db973b84ca295da754c29c9c9a4,
title = "Evolving parameterised policies for stochastic constraint programming",
abstract = "Stochastic Constraint Programming is an extension of Constraint Programming for modelling and solving combinatorial problems involving uncertainty. A solution to such a problem is a policy tree that specifies decision variable assignments in each scenario. Several solution methods have been proposed but none seems practical for large multi-stage problems. We propose an incomplete approach: specifying a policy tree indirectly by a parameterised function, whose parameter values are found by evolutionary search. On some problems this method is orders of magnitude faster than a state-of-the-art scenario-based approach, and it also provides a very compact representation of policy trees.",
author = "S. Prestwich and S.A. Tarim and R. Rossi and B. Hnich",
year = "2009",
month = sep,
day = "19",
doi = "10.1007/978-3-642-04244-7_53",
language = "English",
volume = "5732 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "684--691",
editor = "Gent, {Ian P. }",
booktitle = "Principles and Practice of Constraint Programming - CP 2009",
address = "United Kingdom",
}