EASI (expert advisor for system identification): a prototype package for linear and nonlinear systems

System identification has many aspects which entail experienced and judgemental decisions, as well as very detailed numerical algorithms. Thus, to provide skilled identification requires a mixture of mathematics, heuristics and real process knowledge in addition to the methodology involved in these topics, especially in practical situations. The paper concerns the encoding of such multi-faceted knowledge into an expert system environment. An expert system advice prototype has been constructed to guide the user throughout different stages of the system identification procedure in an attempt to make it as formal as possible, despite its informal nature. The knowledge base being built covers linear, nonlinear, single variable and, to some extent, multivariable techniques