Objective

The project began within the Industry 4.0 framework with the aim of making it easier for small and medium-sized enterprises to adopt the concept and supporting them through implementation.

At Glam we had been rolling out our software at companies for years. Our experience in operations allowed us to identify unmet needs related to operational planning.

The technology used until then worked well as long as the number of rules and resources to consider remained limited; as the variables to control grew, scheduling became cumbersome and inefficient and held back progress.

We needed to take a further step and look for a solution outside our traditional scope, and at that point using artificial intelligence algorithms opened the door to tackling planning requirements we had been unable to resolve before—and, in turn, to deliver extra value to our clients.

The project

We were in a period when artificial intelligence was not yet as widespread as it is today, so proposing its use in operational planning was a major challenge. Specifically, the goal was to deliver a project that could offer companies an AI-driven planning environment.

To make adoption straightforward, the project was not only about building the corresponding AI model; it also required a clean data structure and a technology platform that eased rollout and integration. We therefore chose to create new operations software on a microservices-based platform that could run stand-alone and, where the customer already had working operations software, could also act as a bridge between that system and the planning algorithm.

A project of this scale was neither easy nor cheap, technically or financially, so we pursued partnerships in two directions. On one hand we engaged Eurecat as a technology partner to help define the artificial intelligence model and implement it. On the other, we sought funding through CDTI.

The process

As agreed with CDTI, the project was split into three milestones. In the first we developed a new microservices-based technology platform on which to build a software foundation that could connect to the artificial intelligence algorithm. We designed and implemented a new, modern framework with the base needed to build out the project.

In the second milestone, together with Eurecat, we designed the structure of the planning algorithm with its set of rules and constraints. The algorithm started from a scenario defined by products, manufacturing routings, stock levels, resources and requirements and, through a cost-based weighting system, had to determine the distribution of production orders that met the needs at minimum cost.

Delivering this required defining which data model the algorithm would use, as well as the rules and constraints it had to respect. Those structures had to be developed on the new platform so they could be maintained while Eurecat began work on the artificial intelligence algorithm.

Finally, in the third milestone we consolidated all the work by presenting the platform with the described capabilities and integration with the planning algorithm.

The outcome

Once the project was complete we had a new software foundation on which to build complex operations systems using a modern, advanced technology platform and embedding artificial intelligence algorithms to address highly demanding planning requirements.

This new platform has let us grow as a company and adopt cutting-edge technologies so we can take on larger, more complex and highly scalable projects.