You are here


Constraint Programming For Data Mining Material

Sequential Pattern mining to solve Sequential Pattern Mining Problem with/without time constraint

[PPIC][1] : Find all frequents sequences. Additional constraints are available as well. *Software + Code + Benchmarks

Benchmarks SS-VRPTW-CR

These are the generated test instances for the Static and Stochasic VRPTW with both random Customers and Reveal times (available here).

Uttf instances

On this page, we provide the uttf instances used in the paper “Efficient Filtering for the Unary Resource with Family-based Transition Times that has been submitted at the CP2016 conference.

These instances represent Job-Shop problems with sequence-dependent family-based transition times.

Scalable Time Table Constraint


1) Minizinc Model
2) Large Instances
3) Oscar Source Code of Scalable Cumulative, Synchronized Sweep

Benchmarks for the Dynamic and Stochastic Vehicle Routing Problem with Time Windows

This page contains the benchmarks used in the paper A Multistage Stochastic Programming Approach to the Dynamic and Stochastic VRPTW, Saint-Guillain et al., 2015 (available here).

Table Constraint Propagators

Comet source code for the propagators defined in the Thesis “Propagators for Table Constraints” (Jean-Baptiste Mairy) and state-of-the art propagator re-implementations. Also, python source code for the statistical treatment of the experimental data presented in the thesis. All the code is published under GPLv3 licence.


Benchmarks for The Unary Resource with Transition Times

On this page, we provide the instances used in the paper “The Unary Resource with Transition Times” that has been submitted at the CP2015 conference.

These instances represent Job-Shop problems generated from existing Taillard instances (…) to which have been added transition times.

Segment Routing

This page contains the artificial instances used for our experiments on TE with segment routing.

Discrete Lot Sizing Problem

The Discrete Lot Sizing problem considered here is a multi-item, single machine problem with capacity of production limited to one per period. There are storage costs and sequence-dependent changeover costs, respecting the triangle inequality. The changeover cost q^{i, j} is induced when passing from the production of item i to another one j with q^{i, i} = 0 for all i. Backlogging is not allowed, each order has to be produced before the corresponding demand time,