Genetic algorithms and neural networks

 

algoritmos sup art

 

             This project, developed by the ICT Department of the CTCR, aims at developing algorithms to assist and support the footwear design process.

On the one hand, the final product shall be elaborated through the use of generative design, a method where the final result is elaborated by parameterized components. Through genetic algorithms inspired in the evolution of species, the different components that will make up the generative shoe design will be created randomly. Genetic algorithms are currently also being used to solve problems on the optimization of variables (such as, improve the planning of manufacturing process or ensure the efficacy and efficiency of delivery routes).

On the other hand, algorithms inspired in the human brain and neural networks have been developed to solve complex nonlinear problems, such as the classification of objects or signal analysis. Moreover, these algorithms have the capability of learning and improving their results through analyzing resolved cases previously provided by human experts. With these algorithms, complex marketing strategies can be developed to analyze the brand’s reputation or predict certain scenarios.
The practical applicability of such algorithms is mainly as a tool for the analysis and assistance of footwear designs by generating randomized genetic design and analyzing their sales expectations.