Evolution of Petrinet Modeled Knowledge Intensive Processes

The domain of Knowledge Intensive Process Improvement (Knowiπ) will be considered in this paper. For the improvement of knowledge intensive processes (KnowiP) we suggest Genetic Algorithms (GAs). The representation needed by the GA depends on the modelling language of KnowiPs. We will use Petrinets for the process modelling and then use the incidence matrix of the corresponding Petrinet as problem representation. The processes are evaluated according to performance metrics adapted to knowledge intensive processes. Two particular GAs are applied to improve a special class of KnowiPs, one single objective GA and one multi objective GA, namely the Multi-Sexual GA (MSGA) multiple criteria simultaneously. 1 Knowledge Intensive Process Improvement (Knowiπ) Business processes can be found in any organisation. Today, they are more and more based on required knowledge and supposed to be flexible and adaptable to the changing surrounding conditions as the economic market conditions change rapidly. A BUSINESS PROCESS is a sequence of activities aiming at the creation of one or more products or services with a value for a customer. It is started and finished by one or more events. As it proceeds in an organisation there is an underlying organisational structure [RvHS04]. Furthermore, the activities usually need one or more resources (like people, processors, data, software, etc.) that belong to predefined resource classes. In this paper we want to consider special processes, namely knowledge intensive business processes or simply knowledge intensive processes, as already defined in [GW04]: A process is a KNOWLEDGE INTENSIVE (BUSINESS) PROCESS (KNOWIP) if its value can only be created through the fulfillment of the knowledge requirements of the process participants. Typically knowledge flows and knowledge transfers between media and persons are necessary to achieve a successful process completion. Business processes can be structured, semistructured and unstructured. Structured processes are completely predefined. There are fixed and non changeable rules for the execution of every activity. Semistructured processes contain structured parts and non structured parts. Unstructured processes are completely unpredictable.