News
August 01, 2009

Binary domain is avaiable in beta version:

A dedicated Binary domain has been created on Choco 2.0.1 (beta version). It is not consider stable as we run test on it, but it can be used (methods are fully implemented).

Read More


Links

Search



 

Welcome to ChocoLib!
Some info here

C
hoco is a library implementing the basic tools for PPC management areas, propagation of constraints, and procedures for local and global search.

This library has been implemented under the project OCRE, whose goal is provide a tool Constraints on Research and Education (OCRE).
This project brings together teams from the Ecole des Mines de Nantes, Toulouse Center of ONERA, LIRMM, University of Montpellier and Bouygues DTN, was born from the observation that academic research into PPC needed many experiments, but they were fragmented so far, for lack of a common tool.

We have thus decided to build a platform to take advantage of our mutual software developments and have a unified framework for the experimental comparisons of algorithms.

This tool also intended to serve as support for the teaching of the PPC. CHOCO is the core of this platform. It is free software, developed under the GPL (the Free Software Foundation):

its sources are available to enable all to understand how, if necessary, improve it, and of to benefit the entire community of users. Because of its position as the base layer of the platform, CHOCO is subject to sometimes conflicting requirements. It must be:

  1. Generic, so that the platform OCRE can easily accommodate new components, extending for example, the spectrum of the tool to other areas of constraints, while remaining compatible with the kernel CHOCO.

  2. Small and simple, so that every participant can understand OCRE the kernel to extend it. The simplicity also allows the use for teaching purposes.

  3. Efficient, so that algorithms that would be assessed on the platform OCRE are penalized by an additional systematic due to its core.

From a functional point of view, CHOCO offers nothing new compared to existing products on the market, nor a new global constraint, or a new type of search tree: we are pleased to offer basic utilities that everyone must re-implement a system for manufacturing constraints.

A researcher wishing to make a CPD system for evaluating an algorithm now has two difficulties before them:

not only must it from scratch and implement many pieces that have nothing to do with the topic he wants to deal in addition, he will find little help in the scientific literature since the subjects of practical implementation has so far been regarded as minors by the scientific community of CSF.

Both entry barriers to experimentation led us to mount the project OCRE and incidentally to be presented in this paper the design decisions we have taken to CHOCO. This article is an account of our practical implementation, we hope that it will be useful to other researchers.