A single universal input layer sitting in front of four independent solver engines — CVRP, VRPTW, MDVRP, and MDVRPTW. It handles file parsing and ZIP extraction consistently across all modules, while each module's problem-specific logic and feasibility rules remain fully independent.
Upload directly, or as a ZIP archive. When a ZIP contains more than one candidate file, the Runtime selects one automatically using the priority order below.
| Priority | Format | Description |
|---|---|---|
| 1 | .csv | Unified tabular format — the primary format for real-world operational data |
| 2 | .xlsx / .xls | Business/Excel format |
| 3 | .vrp | Classic CVRPLIB-style benchmark format |
| 4 | .txt | Classic Solomon-style benchmark format |
| 5 | (no extension) | Legacy benchmark instances with no file suffix (e.g. "pr10") |
One instance per request. The Runtime does not run batch/folder jobs — each upload is solved individually.
If the uploaded data doesn't structurally match the selected module, the Runtime rejects it immediately with a clear message — instead of silently mis-parsing the file or consuming credits on a meaningless result.
The input layer is shared. The problem-solving logic underneath is not — each module has its own parser, its own feasibility rules, and its own solver.
Capacitated Vehicle Routing. Single depot, vehicle capacity constraints only.
Adds customer time-window and service-time constraints on top of capacity.
Multi-Depot routing — requires at least two depots explicitly marked in the data.
Multi-Depot routing combined with time-window and route-duration constraints.
Every run returns a full .sol report directly in the browser — unchanged from the original engine output — plus two additional downloadable formats built on top of it.
The original engine report, byte-for-byte unchanged.
Business-readable workbook with Summary, Fleet, Routes, and (when supplied) Baseline Comparison sheets.
Structured output, one row per stop — ready to feed into another system.
A canonical Unified Output object used internally to generate .xlsx/.csv, reserved for a future API/middleware — not offered as a direct download in this phase.
Before running a job, you can optionally provide a baseline to compare against — either typed in directly or extracted from an existing routing report.
Type in the company's existing cost, distance, time, and vehicle count directly.
Only scalar metrics (cost, distance, time, vehicles) are extracted. The baseline's own route sequence is never used as GSL input.
Upload a file directly, or try a live demo dataset with no account required.
Launch GSL Solver View Research & Benchmarks