Case studies

If you've ever assembled your own hardwood floors, you know that they must be made up of a variety of boards of different lengths. Several variables come into play when assembling these boxes of boards, such as inventory availability, board length, customer expectations, or surface area to be covered.

Mirage had already automated a part of their factory to facilitate the design and assembly of the board boxes (nesting). However, production planning and robot management were still done manually and required expertise and exhaustive work from employees who worked with Excel files and built sketches by hand.

As a result of this inflexible planning, Mirage often ended up with surpluses or stockouts and had to address these issues within very tight deadlines.

The objective was to fully automate the factory nesting process. As a solution, we opted for an operations research method that uses a set of data and variables to reach the best possible outcome. In the case of Mirage, the idea was to obtain a perfectly balanced combination of boards through combinatorial optimization (a branch of operations research).

They developed an algorithm that allows for the most heterogeneous combination possible, using the available material perfectly. Like a giant game of Tetris, the robot now distributes the boards in the correct order and in a nearly random way so that no box sent to the customer is identical to another. In this way, Mirage can also optimize the amount of wood per package so that there is almost no remaining stock.

The project was definitely a success for the company. Deployed in just one and a half months, it has already demonstrated excellent results, proving that artificial intelligence can sometimes be deployed quickly and effectively in a factory.

The creation of an algorithm to completely optimize and automate production has allowed Mirage to minimize losses, maximize operation times, and use existing robots to their full potential. The boxes are varied and balanced, recipes are more optimal, quality is increased, and the assembly method is more efficient.