New release: LocalSolver 10.5
We are pleased to announce the release of LocalSolver 10.5, which comes with many new features and performance improvements. Our colleague Léa from R&D gives some details in the following video.
Here is the summary of the main novelties coming with LocalSolver 10.5.
Vehicle Routing Problems
Vehicle routing problems have never been solved faster:
- TSP – Gap to best known solutions <4% for instances with up to 14,000 customers within 4 minutes of running time
- CVRP – Gap to best known solutions <2% for instances with up to 1,000 customers within 4 minutes of running time
- CVRPTW – Average gap to best known solutions 4.6% with up to 1,000 customers within 4 minutes of running time
- PDPTW – Average gap to the best known solutions 7.9% with up to 1,000 customers within 4 minutes of running time
Scheduling & Packing Problems
Packing problems up to 10,000 items optimized in minutes and scheduling problems up to 10,000 tasks optimized in minutes:
- Bin Packing Problem – Average gap to best known solutions 0.7% for instances with up to 6,000 items within 1 minute of running time
- Assembly Line Balancing Problem – Average gap to best known solutions 0.2% for instances with 1,000 tasks within 2 minutes of running time. For 71% of instances, LocalSolver strictly outperforms CPLEX CPO
- Flexible Job Shop Problem – Average gap to best known solutions 2.2% for instances with up to 400 tasks within 1 minute of running time
Portfolio optimization problems
LocalSolver now solver of choice for portfolio optimization problems thanks to novel convex optimization techniques:
- General improvements in Mixed-Integer Non-Linear Programming (MINLP), more MINLPLib instances solved to global optimality with faster solving times
- In the case of Convex Mixed-Integer Quadratic Problems (MIQP), LocalSolver is now significantly faster
- For example, the 60 instances of the Small-Investor Mean-Variance Portfolio Optimization Problem from MINLPLib are now solved to global optimality in seconds. This problem extends the standard Markowitz Mean-Variance Portfolio Optimization Problem with integrity constraints imposed on decision variables
LocalSolver Cloud
LocalSolver Cloud allows your models to be anonymized, encrypted, sent, and optimized on powerful computers on the cloud at no extra cost.
New modeling features
Set-based modeling is enhanced with the COVER operator to model problems where an item must appear in at least one set or list, but possibly more.
For example: Split Delivery Vehicle Routing Problems (SDVRP)
Modeling is made easier with the FIND operator retrieving which collection contains a given item.
For example: Flexible Scheduling, Assembly Line Balancing
New examples
New examples available in Example Tour to help you get started:
We are at your disposal to accompany you in the discovery of LocalSolver 10.5. Just ask for your free trial license by registering here. In the meantime, don’t hesitate to contact us for any further information or support.