The Ubiquity Generator Framework: 7 Years of Progress in Parallelizing Branch-and-Bound

  title={The Ubiquity Generator Framework: 7 Years of Progress in Parallelizing Branch-and-Bound},
  author={Yuji Shinano},
Mixed integer linear programming (MILP) is a general form to model combinatorial optimization problems and has many industrial applications. The performance of MILP solvers has improved tremendously in the last two decades and these solvers have been used to solve many real-word problems. However, against the backdrop of modern computer technology, parallelization is of pivotal importance. In this way, ParaSCIP is the most successful parallel MILP solver in terms of solving previously… 

An Easy Way to Build Parallel State-of-the-art Combinatorial Optimization Problem Solvers: A Computational Study on Solving Steiner Tree Problems and Mixed Integer Semidefinite Programs by using ug[SCIP-*,*]-Libraries

A computational study for solving Steiner tree problems and mixed integer semidefinite programs in parallel and presents results for previously unsolvable instances from the well-known PUC benchmark set, widely regarded as the most difficult Steiners tree test set in the literature.

Solving Previously Unsolved MIP Instances with ParaSCIP on Supercomputers by using up to 80,000 Cores

The basic parallelization mechanism of ParaSCIP is described, improvements of the dynamic load balancing and novel techniques to exploit the power of parallelization for MIP solving are described.

Building Optimal Steiner Trees on Supercomputers by Using up to 43, 000 Cores

This paper describes an updated version of ug [SCIP-Jack, MPI], especially branching on constrains and a customized racing ramp-up, and the different stages of the solution process on a supercomputer are described in detail.

The SCIP Optimization Suite 5.0

New features and enhanced algorithms made available in version 5.0 of the SCIP Optimization Suite, in particular for the LP solver SoPlex, the Steiner tree solver SCIP-Jack, the MISDP solverSCIP-SDP, and the parallelization framework UG are described.

Linearization and parallelization schemes for convex mixed-integer nonlinear optimization

These parallelization methods are compared to alternate approaches that exploit parallelism in existing commercial MILP solvers and the latter approaches are seen to perform better thus highlighting the importance of MILP techniques.

MIPLIB 2017: data-driven compilation of the 6th mixed-integer programming library

For the first time, these sets were compiled using a data-driven selection process supported by the solution of a sequence of mixed integer optimization problems, which encode requirements on diversity and balancedness with respect to instance features and performance data.

Automated airport staff scheduling at Swissport International Ltd.

My first year as IFORS President and about what the authors have achieved in these months are thought about, and each of us has started understanding the current situation, creating contacts, initiating activities and making plans for the future.

The Impact and Implications of Optimization

  • J. Kallrath
  • Business Optimization Using Mathematical Programming
  • 2021



ParaSCIP: A Parallel Extension of SCIP

ParaSCIP is presented, an extension of SCIP, which realizes a parallelization on a distributed memory computing environment and was able to solve two previously unsolved instances from MIPLIB2003, a standard test set library for MIP solvers.

A First Implementation of ParaXpress: Combining Internal and External Parallelization to Solve MIPs on Supercomputers

This paper presents a first implementation of ParaXpress, a distributed memory parallelization of the powerful commercial MIP solver FICO Xpress, and provides computational experiments to address the question how to balance the internal Xpress parallelization and the external parallelization by UG against each other.

Parallel PIPS-SBB: multi-level parallelism for stochastic mixed-integer programs

Two different parallelizations of Branch & Bound (B&B) are presented, implementing both as extensions of PIPS-SBB, thus adding an additional layer of parallelism.

Effectiveness of Parallelizing the ILOG-CPLEX Mixed Integer Optimizer in the PUBB2 Framework

A new method of parallelizing a MIP (Mixed Integer Programming) solver is introduced, different from a standard implementation that constructs a parallel branch-and-cut algorithm from scratch (except using an LP solver).

PIPS-SBB: A Parallel Distributed-Memory Branch-and-Bound Algorithm for Stochastic Mixed-Integer Programs

PIPS-SBB is an exact distributed-memory parallel stochastic MIP solver that takes advantage of parallelism at multiple levels of the optimization process and shows promising results on instances from the SIPLIB benchmark.

Solving Hard MIPLIB2003 Problems with ParaSCIP on Supercomputers: An Update

For the first time, computational results of single runs for two open MIPLIB2003 instances could be solved by ParaSCIP in more than ten consecutive runs, restarting from checkpointing files are presented.

BOB : a Unified Platform for Implementing Branch-and-Bound like Algorithms

The BOB library is founded on the notion of global priority queue which makes the parallelization methods independent from the applications, and vice-versa, and this library has the double goal of allowing on the one hand the Combinatorial Optimization community to implement their applications without worrying about the architecture of the machines and beneeting the advantages provided by parallelism.

Distributed Domain Propagation

Portfolio parallelization is an approach that runs several solver instances in parallel and terminates when one of them succeeds in solving the problem. Despite its simplicity, portfolio

ParaLEX: A Parallel Extension for the CPLEX Mixed Integer Optimizer

This paper introduces ParaLEX which realizes a master-worker parallelization specialized for the solver on a PC cluster using MPI and shows that Para LEX is highly effective in accelerating thesolver for hard problem instances.

Solving Open MIP Instances with ParaSCIP on Supercomputers Using up to 80,000 Cores

The basic parallelization mechanism of ParaSCIP is described, improvements of the dynamic load balancing and novel techniques to exploit the power of parallelization for MIP solving are described.