Production Scheduler

What it solves

Production scheduling often gets stuck on the same details: orders arrive as spreadsheets, due dates and priorities differ, processes can run on different machines, and shifts plus maintenance windows change when work can actually start. production-scheduler turns this into a repeatable scheduling flow: start from an order CSV, choose an algorithm, produce an HTML Gantt chart, and optionally generate a simulation page.

How the skill works

• Inputs include order_id, product, quantity, arrival_time, due_date, and priority in CSV form.
• It supports three algorithms: fcfs by arrival order, greedy by selecting the least-loaded machine at each step, and aco by iteratively reducing machine load imbalance.
• Shifts can be set with --shift to 24h, day, 2shift, or 3shift; working days can be customized with --work-days. Machine-level maintenance windows can also be skipped automatically.
• Default process and machine mappings live in scripts/scheduler.py, for example P1 cutting can use M1 and M2, while P3 welding can use M5 and M6.
• Adding --sim produces an interactive simulation page for workpiece flow, equipment status, WIP counts, and playback speed control.

Boundaries and notes

This is a good fit for small-to-medium shops with a relatively fixed process structure. If order fields, machine topology, or routing change frequently, the CSV schema and default process configuration need to be updated together. ACO parameters such as n_ants, n_iterations, alpha, and beta affect runtime and result stability, so they should be tuned against real constraints.