Learn Mini-CP

Technical Report

The complete architecture of mini-cp is described this document.

Slides and teaching material

Slides

Slides from the course INGI2365 given by Pierre Schaus @UCLouvain:

Learning outcomes by studying mini-cp:

From a state and inference prospective, specific learning outcomes include:

  • Trailing and state reversion
  • Domain and variable implementation – Propagation queue
  • Arithmetic Constraints
  • Logical Constraints
  • Reified Constraints
  • Global Constraints (including for scheduling)
  • Views

From a search prospective, the outcomes include:

  • Backtracking algorithms and depth first search
  • Branch and Bound for Constraint Optimization
  • Incremental Computation
  • Variable and Value Heuristics implementation
  • Searching with phases
  • Large Neighborhood Search

While, from a modeling perspective, the outcomes include:

  • Redundant constraints
  • Bad smells and good smells: model preferably with element constraints instead of 0/1 variables
  • Breaking symmetries
  • Scheduling: producer consumer problems, etc.
  • Design problem specific heuristics and search

Mini-CP XCSP3 Mini Solver

We provide under the form of a student project the possibility to participate to the XCSP3 MiniSolver Competition with Mini-CP. All the interfacing with XCSP3 tools and the parsing of XCSP3 format is done for you. You can focus on the only interesting part: make your solver as efficient as possible.