Obiettivo Today, it is becoming increasingly difficult to sell industrial machines without reliable customer support. The "cost of ownership" of long lasting equipment often far exceeds its initial value. Buyers increasingly see maintenance costs as the deciding criterion. By increasing the quality of customer support with help-desk software and by bundling diagnostic software together with his equipment, a supplier will score points against competitors and gain market share. European manufacturers of equipment have a long tradition of tracking fault histories, but they rarely take full advantage of this. The leading edge technology of Case-Based Reasoning (CBR) helps discover decision knowledge in historical databases to:- Improve after-sale support (telephone hotline) with help-desk software.- Develop diagnosis and fault analysis decision support systems.- Update regularly troubleshooting manuals from observed faults.- Capture and reuse the experience of the most talented maintenance specialists, transfer expertise amongst personnel and build a corporate memory.CBR helps engineers relate current problems to past experiences. A new problem is solved by finding similar cases and adapting the solution that worked in the past to the current problem. CBR optimises the number of tests required for troubleshooting, reduces the cost of repair and minimises downtime of mission critical equipment. The technology has an intuitive appeal for support engineers and adapts to their mode of reasoning: "Have I ever seen a similar problem before and if so, what did I do about it?". By not requiring troubleshooting specialists to describe their know-how as logical rules, CBR overcomes what has, historically, been the main stumbling block in building rule-based expert systems. APPLICUS aims at demonstrating the suitability of European CBR technology for a real customer support problem of robots, at refining the methodology for building such applications, and at identifying relevant measurable parameters to quantify return on investment from the manufacturer's perspective in order to promote the technology by demonstrating its usefulness.The consortium involves:- AcknoSoft, a French company that specialises in Case-Based Reasoning technology and that was the prime contractor of the ESPRIT project INRECA (P6322).- BSR Consulting, a German company that delivers turnkey solutions with advanced computer technologies to its clients. BSR was a partner in the German project FABEL and has acquired experience in different aspects of CBR.- The after-sale division of SEPRO Robotique, a French company that manufactures robots for plastic injection moulding. They have installed a CBR help-desk in collaboration with AcknoSoft. SEPRO will identify relevant parameters to measure to quantify the return on investment from using the technology, with a view to deploy it world-wide to its different representatives. This will demonstrate the usefulness of CBR technologies for SMEs. Campo scientifico natural sciencescomputer and information sciencessoftwarenatural sciencescomputer and information sciencesdatabaseshumanitieshistory and archaeologyhistorynatural sciencescomputer and information sciencesartificial intelligenceexpert systems Programma(i) FP4-ESPRIT 4 - Specific research and technological development programme in the field of information technologies, 1994-1998 Argomento(i) 1.6 - Emerging Software Technologies Invito a presentare proposte Data not available Meccanismo di finanziamento EST - Marie Curie actions-Early-stage Training Coordinatore Acknosoft Contributo UE Nessun dato Indirizzo Rue Du Dessous Des Berges 58A 75013 Paris Francia Mostra sulla mappa Costo totale Nessun dato Partecipanti (2) Classifica in ordine alfabetico Classifica per Contributo UE Espandi tutto Riduci tutto BSR CONSULTING Germania Contributo UE Nessun dato Indirizzo WIRTSTRAßE 38 81539 München Mostra sulla mappa Costo totale Nessun dato Sepro Robotique Francia Contributo UE Nessun dato Indirizzo Boulevard Des Etats-Unis 44 85000 La Roche sur Yon Mostra sulla mappa Costo totale Nessun dato