Hexaly’s Post

🚀 We are pleased to announce the release of Hexaly 12.5, which comes with many new features and performance improvements. • Capacitated Vehicle Routing Problem with Time Windows (CVRPTW): average gap of 1.8% in 1 minute for instances with up to 1,000 points • Pickup and Delivery Problem with Time Windows (PDPTW): average gap of 0.8% in 1 minute for instances with up to 1,000 points • Scheduling Problems with Production and Consumption of Resources: average gap of 3.0% in 1 minute on Kone’s benchmarks • Resource Constrained Project Scheduling Problem (RCPSP): improvement of 72 world records from the literature, in addition to the 121 improved by Hexaly 12.0 • Bin Packing Problem: average gap of 1.2% in 10 seconds on instances with 5,000 items and near-optimal solutions in minutes for instances involving 100,000 items 🔗 Want to learn more? Check out the link below: https://lnkd.in/eg-vSav2 #optimization #operationsresearch

Hexaly against Gurobi and OR-Tools on the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW): ➡️ Hexaly finds solutions close to best known solutions for instances with up to 1,000 customers. ➡️ Gurobi struggles to find feasible solutions for instances with more than 400 customers. ➡️ OR-Tools finds solutions with a gap greater than 10% for instances with more than 600 customers. 🔗 Check the benchmark here: https://lnkd.in/e5sGgZWC #optimization #operationsresearch

We're thrilled to announce that Hexaly will host a stand at the 33rd European Conference on Operational Research, EURO 2024, at the Technical University of Denmark from June 30 to July 3, 2024. Visit our booth to discover the latest features and applications of Hexaly 12.5, meet our team, and explore our job opportunities. Feel free to attend the presentations given by our team to learn more about Hexaly: 🔸Hexaly, a new kind of global optimization solver by Fred Gardi 🔸Hexaly Studio: a no-code platform for mathematical optimization by Thierry Benoist 🔸Hexaly Studio: no-code modeler for routing and scheduling problems by Lucas Ligny 🔸Automatic model decomposition in Hexaly Optimizer by Julien Darlay 🔸Operational planning of medical vehicle routes by Nicolas Blandamour 🔸Disjunctive scheduling using interval decision variables with Hexaly Optimizer by Léa Blaise 🔗 For more details: https://lnkd.in/efHjXuKX

Hexaly against Gurobi on the Traveling Salesman Problem (TSP): On the Traveling Salesman Problem, Hexaly reaches an average optimality gap of 0.3% in 1 minute of running time on the TSPLIB research benchmark with instances up to 10,000 cities. Gurobi fails to find decent solutions for instances with more than 300 customers. 🔗 Check the benchmark here: https://lnkd.in/eqFEPzns #optimization #operationsresearch

Hexaly against Gurobi, OR-Tools, jsprit (GraphHopper), and OptaPlanner on the Capacitated Vehicle Routing Problem (CVRP): ➡️ Hexaly finds solutions close to the state of the art in 1 minute, even for instances with 1,000 customers. ➡️ Despite specializing in solving Vehicle Routing problems, OR-Tools delivers poor-quality solutions (> 5% gap), even for medium-sized instances; OptaPlanner and jsprit perform worse (>10% gap). ➡️ Gurobi fails to find decent solutions in 1 minute, even for small-sized instances; additional experiments show the results remain the same with 1 hour of computation. 🔗 Check the benchmark here: https://lnkd.in/etzyrd-8 #optimization #operationsresearch

Hexaly is thrilled to sponsor the 15th Metaheuristics International Conference this week in Lorient, France. Visit our booth to discover the latest features and applications of Hexaly 12.5 and explore our job opportunities! #optimization #operationsresearch

Hexaly against Gurobi on the Bin Packing Problem: For the 240 BPPLIB instances commonly used by the research, Hexaly achieved an average gap of 0.2% in 1 minute of running time on a standard server (8 cores, 3.8GHz). After 1 minute, the *worst* gap observed for Hexaly is 0.6%; after 10 minutes, the *worst* gap is 0.3%. After 1 minute, the *average* gap observed for Gurobi is 44%; after 10 minutes, the *average* gap is 11%. Hexaly can solve Bin Packing problems involving 1,000,000 items in minutes of running time. On the other hand, traditional MILP solvers like Gurobi struggle to find quality solutions for instances with 1,000 items. 🔗 Check the benchmark here: https://lnkd.in/eCgEWpaF #optimization #operationsresearch

How does Hexaly perform against CP Optimizer, OR-Tools, Gurobi, and Cplex on the Job Shop Scheduling Problem (JSSP)? In this benchmark, we focus on large-scale instances of the Job Shop Scheduling Problem (JSSP): 1,000 jobs and 1,000 machines for 1,000,000 activities. Hexaly offers an innovative modeling approach based on interval and list variables, making the Job Shop Scheduling Problem (JSSP) modeling compact and straightforward. Then, Hexaly greatly outperforms traditional solvers and improves the research's best known solutions by an average of 7.4% within 60 seconds. 🔗 Dive deeper into the details, explore our blog post: https://lnkd.in/enurjcaM #optimization #operationsresearch #datascience

🏆 New Records for the Inventory Routing Problem (IRP) Here is remarkable progress in addressing the Inventory Routing Problem (IRP), a variant of the Vehicle Routing Problem (VRP) involving inventory management: Hexaly has enhanced the best solution known in the research for 4 out of 11 instances of the IRP 🔗 Dive deeper into Hexaly's achievements by reading the full article on our website: https://lnkd.in/e2QySM6j #optimization #operationsresearch