Clustering alternative product formulations using graphs

Autor: Javier A. Arrieta-Escobar, Fernando P. Bernardo
Rok vydání: 2019
Předmět:
DOI: 10.1016/b978-0-12-818634-3.50086-2
Popis: This work proposes the use of graph partitioning tools to identify clusters in a possible large space of plausible product formulations, with each cluster representing a significantly different design alternative that deserve to be tested. The partitioning tools are used to find clusters in a large set of previously generated alternatives, and are also incorporated into a mixed-integer programming problem whose solution generates clustered alternatives (within a feasible space defined by available models and heuristics). An example of a cosmetic emulsion is provided.
Databáze: OpenAIRE