Optimize over 1,200 daily orders relative to an extremely varied portfolio of products in terms of types of steel, measurements, mechanical processing, thermal treatments and coatings.

Productive constraints also had to be considered, including: varied external machining, equipment capacity, the grouping of orders to launch a work cycle, organizational constraints, and production sequence limitations.


The client was using two tools to plan operations: a management system that assumed infinite production capacity for the plant, and a traditional, generic scheduling system that didn’t reflect the technical limits of production, scheduling each center locally.
Programming and scheduling were thus very cumbersome in terms of time, and quite inefficient in terms of results.


We created a tool that fully optimizes the production workflow, while taking technical, time and capacity restraints into consideration. As such, the daily planning and monitoring of production were simplified, through the creation of a schedule that minimizes waste (both time and materials) to boost earnings.

This tool makes it possible to support each manufacturing department in optimizing its workflow on a short-term timeframe (generally a few days) and make dynamic interaction possible for reprogramming and the creation of what-if analyses (useful to simulate the unavailability of equipment and raw materials, the availability of back-up/alternative equipment, urgent requests, etc.).


Through the system developed, the company can now plan and schedule new orders in such a way that avoids overloading its facilities. Delays have become negligible for a few product categories, reducing them notably for others. The flow of materials in the department has improved, as has the OTD (on time delivery) indicator, which means better service is provided to customers and improved cash flows. Management costs were reduced in relation to scheduling activities through the creation of a tool to share expertise and knowledge, the company’s most important asset. The company’s ability to successfully cope with unforeseen events has improved through the option to simulate different scenarios and understand their impact on system capacity.


Late orders



Average delay days



Work in process



Orders late more than 2 weeks


Sviluppi in corso

An in-house R&D project is currently in progress, aimed at the implementation of an artificial intelligence system as an alternative to classic optimization. The method currently being tested is based on multiagent deep reinforcement learning (MDRL), which, while maintaining the same product delivery date, makes it possible to look ahead, seeking local optimums as solutions that improve the current planning methods in relation to different scenarios.
This would make it possible to optimize unplanned event management, reducing the time it takes to calculate solutions and reacting in a data-driven way to the sudden changes that may occur during production.