A successful collaboration between CETIC and Uppsala university

A successful collaboration between CETIC and Uppsala university

MiniZinc Challenge 2016

The OscaR.cbls optimisation engine developed by CETIC and adapted by Uppsala has been ranked first in its category "local search engines" during an international competition between solvers.

Date: 2 December 2016


Algorithmic and combinatorial Optimisation 

Domaine: Software industry 

Asset: Oscar.CBLS 

About projects

Such competition aim at fostering the development of optimization engines because it offers an opportunity to objectively compare then against a set of standard challenges. Optimization engines are then used in practice to solve real-world problems such as routing and scheduling problems.

An optimization engine relies on a fundamental algorithm. Local search is one of these. Its is able to cope with large problems. CBLS is a way to make local search easy to use and to adapt to specific optimization problems. CBLS stands for "Constraint-Based Local Search".

Since 2012, CETIC has been developing an optimisation engine, based on local search. This engine is part of the OscaR open source framework and is called OscaR.cbls.

Uppsala university in Sweden has developed a tool called OscaR/cbls/MiniZinc that can solve any optimization problem that is specified in a standard language called MiniZinc. This tool transforms MiniZinc problems into the format supported by OscaR.cbls and uses it as a back-end solver.

This tool took part to the annual competition of optimization engines for MiniZinc called the MiniZinc Challenge 2016. OscaR/cbls/MiniZinc was ranked first among competing local search engines. The ranking is based on the quality of the result computed by the engine as well as the speed of the tool to produce it.

View online : http://www.minizinc.org/challenge20...