Identifying validation categories associated with agent-based and system dynamics modelling
The process of identifying validation categories for computer-generated forecasts began by defining search terms focused on either model or method validity in conjunction with agent-based and system dynamics modelling techniques, as shown in Fig. \ref{595505}. In order to identify as many different validation traits for these forecasting techniques as possible, the selected keywords were kept deliberately simple (see point 1 below) to avoid unintentionally excluding records spread across different industrial and commercial domains. By scanning the contents of academic journals for these key words (using the Google Scholarâ„¢ and Web of Knowledge academic search engines) over 200 papers were identified that might provide insight into the different approaches adopted to demonstrate the credibility of forecast results. By then manually reviewing all of the abstracts provided for this initial set of academic papers, a further down-selection took place to focus in on papers that placed greater emphasis on either reviewing methodological developments or on the application of these specific simulation techniques for predicting disruptive changes. In this manner the final set of literature sources to be analysed was identified, as summarised in Table \ref{table:literature_simulation_mapping}. The literature sources recorded here cover a broad range of domains (including agriculture, biology, city planning, ecology, economics, health-care, and transportation, amongst others) in order to identify alternative validation techniques applied within different industries.