Sales support system and manufacturing scheduling
In this project soft computing methods has been applied for complicated manufacturing problems. The research work has taken place in real manufacturing environment at the industrial plant of ABB company in Pitäjänmäki in Helsinki. The target of the research has been a developement of a sales support system for the sales process of synchronous machines and a rough scheduling system for the manufacturing process of them.
Developement work of computer aided manufacturing systems is well-known to be a difficult task because existing information to be included to the system is multiformal. Traditional programming, database, and expert systems are not versatile enough to utilize the existing information. A massive amount of data may exist with a complicated interdependences. Furthermore, the information may be uncertain, imprecise, vague etc.
By applying modern artificial intelligent techniques in the form of soft computing many of the problems which has been unsolved beforehand are now cabable of settlement. The most prominent soft computing techniques are fuzzy logic, neural networks, and genetic algorithm. In the applied research of softcomputing in the production environment of synchronous machines, large problems have been split to sub-problems, which the appropriate soft computing method is applied. Those modular parts is then integrated as compound system which can be seen as a hybrid system of different solution methods as well as the nature of the appearing information.
Sales support system is an utensil for salesman to make resonable offers. In a case of customized product as synchronous machine the automatic configuration and pricing is difficult task by ordinary programming tools. Manufacturing scheduling is well known to be a very complicated problem. Especially in ABB case the complexity is huge. The purpose has not been to solve problem completely this will say in realtime or in high accuracy level. Instead static type of factory simulation is done. Genetic algorithm is used as optimization algorithm and neural networks to model the manufacturing cells.
This project is a subproject of the HYPE project which belongs to TEKES technology program Adaptive and Intelligent Systems Applications.
Tel.: +358-9-470 25149
Fax: +358-9-470 23308
- E Koskimäki,
"A genetic algorithm for capacity analysis and scheduling problem of electric machine manufacturing,"
in Advances in Systems, Signals, Control and Computers, V Bajic, South Africa: IAAMSAD, 1998, s. 212-215.
- E Koskimäki, J Göös, P Kontkanen, P Myllymäki and H Tirri,
"Comparing soft computing methods in prediction of manufacturing data,"
in Lecture notes in artificial intelligence, A Pasqual del Pobil, J Mira and A Moonis, Berlin: Springer, 1998, s. 775-784.
- E Koskimäki and J Göös,
"Electric machine dimensioning by global optimization,"
in Proceedings of the first international conference on knowledge-based intelligent electronic systems, L Jain, Adelaide: IEEE, 1997, s. 308-312.
- E Koskimäki and J Göös,
"Electric machine design by fuzzy fitness,"
in Proceedings of the second nordic workshop on genetic algorithms and their applications, J Alander, Vaasa: Vaasan yliopisto, 1996, s. 237-244.
- J. Göös, E. Koskimäki, and Aarne Halme,
"Utilization of Artificial Intelligent Techniques in Making Offers,"
in Second Nordic Workshop on Genetic Algorithms and their Applications. Vaasa, Finland, 19-23 August 1996.