Design and selection of products via genetic algorithms and neural networks (Q2711710)

In this paper, a new methodology finalized to the design and selection of consumer goods is defined. Generally, design and selection have been kept as two separated processes brought together by the concept of ``mass customization''. ``Mass customization'' is defined in this work as the characterization of consumer goods (products) based on customer's requirements. The problem is looked from the enterprise prospective. That is: given a list of ``feasible'', products, the enterprise must decide which product with which design goes to production. It is assumed that sale forecasts are available and used to estimate the volume of sale of each product in the assigned market segment. Without loss of generality, it is assumed that each product is assigned to a unique market segment. The general approach to solve the ``selection of products'' has been via mathematical programming. NEWLINENEWLINENEWLINEDifferently from the various approaches, the problems considered in the paper are analysed in the framework of genetic algorithms and neural networks. It is proposed a solution strategy that determines for the enterprise the ``best'' design for each of the product brought to market. The enterprise's profit based on revenue estimates for each product in a given segment market is maximized. Moreover, it is assumed that a product has to be offered in the market segment \(k\) with given attribute values.