Engineering Design for the next Millennium: the Challenge of Artificial Intelligence

Much better results in terms of perforniancekost ratio and reliability may be achicvcd if largc machines and structures are replaccd by swarms of smaller ones capablc of co-opcrating on spccificd tasks and negotiating ~ i t h each other what to do next. Armadas of small communication satellites are being dcsigncd to replace largc geostationary units and swarms consisting of hundreds of miniature spacecraft arc to bc sent into thc dccp spacc instcad of a single ship. Closer to earth, swarms of miniature agricultural macl~ncs arc likcly to replace large tractors whilst colonies of co-operating robots arc already making an impact in expcrimcntal manufacturing plants. It is unlikely that all the skills necessary for designing and manufacturing of swarms of intelligent machines will bc available at any one geographical location. Teams of geographically distributcd designers will bc cmploycd communicating over thc Internet using Virtual Dcsign Studio facilities and experimenting in Virtual Laboratories and Virtual Prototping Workshops. Manufacturing will be done in distributed Virtual Factories. Thc changcs in the way engineering products and systems will be designed amount to a genuine paradqyn shift. Kejwords Artificial Intelligence in Engmeering, Multi-Agent Systems, Design of Swarms of Intelligent Machines.