Optimization of Diversity and Distribution of Cable Configuration (OPTDDCC) for Automobile Industry Project

OPT-DDCC: Universidade de Aveiro (Programa Aveiro Norte)/Porto Technical centre of Yazaki Saltano de Portugal, C.E.A. Lda.

The main goal of this complexity optimization project was to produce an optimization and database software designated by DONE – Diversity Optimization on a Network Environment which determines a minimal diversity of configurations from a universe defined by the statistical data about costumer preferences, selling expectations, etc, fulfilling cost limits and other demands like the upper bound on the maximum number of different configurations which are technically possible to produce.

The OPT-DDCC project is one of the three projects shortlist for the Innovation Award APDC/SIEMENS 2006

Abstract: Cars are purchased with a set of active options (airbags, air conditioner, radio car, etc.) which varies from one client to another. A car has an active option connection if it is prepared to include such option as an active one (that is, the car has all the necessary material for connections). In general, each unit is produced with more active option connections than the ones which became active. If the car has an active option connection which not became active, then there is cost (for instance in copper and several other material related with the connections). The global cost of option connections is very high in automobile industry and thus it is crucial to produce cars prepared with a set of active option connections close as possible to the set of active ones. On the other hand, for technical reasons, it is not possible to produce a large variety of different option connection configurations. The implemented system is based in optimization algorithms acting on a mathematical model generated from a database with options, client preferences, configurations, inclusion relation configurations, etc. The optimization problem, the “minimum arc cost sum spanning star forest problem”, is a NP-hard problem and the high dimension of the data implies several problems related with numerical computations, memory management and data manipulation.

This project was developed between January and December 2005 in the facilities of Mathematics Department of Aveiro University.

CEOC Team: