Expert system selection method for SMT placement equipment

Introduction. Until recently, evaluation of Surface Mount Technology (SMT) placement equipment consisted of selecting the machine capable of meeting requirements without regard to price or other factors. Today, several manufacturers offer capable and reliable equipment making additional analysis necessary. No comprehensive evaluation model exists for SMT and mistakes in equipment selection are often made. Common evaluation errors are: limiting analysis to financial factors, use of inappropriate methods, use of incomparable specifications, evaluation with missing information, and overestimation of equipment capacity. Incorrect selection analysis leads to disappointment with equipment capabilities and performance.; Methodology. A selection method has been developed for SMT. The evaluation uses a multi-attribute approach consisting of financial and non-financial factors. Financial evaluation is performed with a modified cost per placement calculation. Non-financial and multi-attribute ratings are determined using an aggregated utility function. Several heuristic rules are applicable to SMT equipment evaluation making traditional computer languages inappropriate. An artificial intelligence language and expert system methodology is used to manage the heuristic rules. The system works in three steps: capability matching, quantifiable performance assessment, and non-monetary performance assessment. The system determines the multi-attribute machine ratings, sorts by that rating, and presents a listing of recommended equipment. The program is in a decision support system format that performs calculations and makes recommendations but leaves the user in control to make the selection.; Results. Equipment recommendations made by the system differ from those of simpler evaluation methods. SMT experts surveyed agree that the system recommendations reflect expertise in the subject domain and that the financial model is a significant improvement to the cost per placement rating commonly used. The system has three significant advantages over current selection methods for SMT. First, the evaluation method is more accurate than current models. Second, knowledge of SMT machine capabilities is available through a system database which eliminates the need for prior research. Third, the system is the first fully functional expert system for equipment selection or justification in any domain.