Method to generate a controllable and adaptable network of models of behaviour templates

The invention relates to a method for generating a method network in the form of networked models of behavior patterns and its optimization in controlled and iterative steps, in particular for problem areas without deterministic solution methods and for methods models with open transition spaces using a MOF-compliant repository on a computer system. As a result, the invention also relates to the generation of software systems from models, in particular for any controllable, regular or customizable sequences of modeled behavior patterns and the adjustable cooperation of methods, agents or other software artifacts. To create the task, automated in different application areas including software production models from process patterns arbitrarily networked,with which, at the level of modeling, elements are automatically generated from which arbitrarily complex aggregations of elements are generated, which in turn can be elements in further aggregates. Such elements produced by the present method are, in particular, patterns, frameworks, software and other model systems, as well as the invention-specific self-adaptive polymorphic agents;in which all elements are described via a four-layer metamodelling in such a way that, firstly, any aggregration mechanisms of the inventive method can be used and, secondly, all elements can be interpreted and manipulated via their model and the modeling technique contained therein;the common method used to generate or regenerate run-time units from new or modified models and integrate them into the current runtime environment.