Flowty

Flowty develops algorithms and solvers designed for multi-commodity flow and large-scale network optimization problems. Its technology addresses complex planning challenges in domains such as logistics, supply chain, scheduling, and infrastructure, focusing on scenarios where multiple origin-destination demands must be routed efficiently through capacitated networks. The platform emphasizes precision, speed, and scalability, offering specialized methods that exploit network structure to deliver faster solutions and clearer decisions compared to general-purpose optimization approaches.

Key features of Flowty include a mixed-integer programming core optimized for network-structured models, decomposition and column generation techniques to handle scaling across many demands, and engines for resource-constrained shortest path problems. The technology supports modeling rich constraints without sacrificing solvability and is purpose-built for large-scale network instances, including resource-constrained and multi-commodity flows. Flowty's research highlights innovations such as tree-based formulations for multi-commodity flow, column generation over source-rooted trees to reduce problem size, and parallel pull labelling for resource-constrained shortest paths, utilizing bucket-level parallelism and vectorized dominance checks.

The tool is positioned for use in real-world optimization scenarios, as illustrated by its solver examples: multi-commodity flow routing, fixed-charge network design, resource-constrained flows, resource-constrained shortest path pricing, vehicle routing with time windows, and set-partitioning assignments. These capabilities target planning problems that require handling capacity limits, fixed costs, time windows, resource constraints, and assignment coverage.

Flowty is available for acquisition, with its algorithms, solvers, and intellectual property offered for purchase. The platform provides documentation and solver access through its website.