Working principle

A concise overview of our optimization approach and edge‑native design.

Overview

Easimplex combines linear/mixed‑integer programming with practical heuristics to deliver fast, explainable plans for real‑world logistics and industrial engineering workloads.

Core models

We formulate assignment, knapsack, cutting‑stock variants as LP/MIP with bounded instance sizes. We use presolve, cutting planes, and branch‑and‑bound tuned for common operations constraints.

Routing heuristics

For VRP with time windows and capacities, we employ constructive methods (sweep, savings), local improvements (2‑opt/3‑opt), and large‑neighborhood search guided by multi‑objective scores.

Edge performance

Workers execute within strict compute budgets. We cap instance sizes, cache reusable subresults, and incrementally improve solutions to maintain low latency on Cloudflare.

Explainability

Plans expose constraint sets, scoring breakdowns, and improving moves so planners can audit and adjust with confidence.

Privacy & security

We minimize retained data, isolate per‑tenant storage, and apply least‑privilege access. Email verification and manual approvals reduce abuse by default.

Back to Algorithms