Inferring control structures from the behaviour of a production system

A control structure for production systems derived from the control structure of a programmed grammar is introduced. This type of control structure is proved to be more general and flexible than the classical control structure based on the concept of a Markov Normal algorithm. A method is discussed for inferring a control structure consistent with the observations of the behaviour of a production system whose rules are known. Finally, some criteria are given for choosing a suitable control structure from the set of control structures actually inferred