The Agent-Based Model

The agent-based implementation of CARIS is characterized by two different classes of agents: (1) the Competitive Environment (CE) and (2) the Competent Actor (CA), the latter representing different categories of actors—such as firms, research groups, research institutions, mediators of innovation, local institutional actors— incorporating the competencies illustrated in Fig. 3. The characterization of each CA is given by the mix of competencies it exhibits and by the actions it performs. Table 2 compares the conceptual model and the agent-based model in terms of actors, actions and outcomes of actions. The main aspects of the agent-based model are presented in Fig. 3.

The First Agent: The Competitive Environment CE

The Agent CE is characterized by a binary string (-1 or 1) of length l, representing the Environmental Regularity (ER) that a Competent Actor CA should discover and match up. The Environmental Regularity is the combination of knowledge and competences that a CA should exhibit in order that a particular market segment can be served with a specific product. Put differently, it represents the proper recipe to satisfy CE requirements. The Agent CE makes three actions:

Fig. 3 The CARIS agent-based model

Table 2 Comparison between conceptual framework and agent-based model

Action 1.1: The generation of Environmental Regularities ER: The Environmental Regularity (ER) changes over time according to a given volatility v; the volatility represents the expiration date of the ER. Once the latter is expired the CE has to generate a new one.

Action 1.2: The generation of Innovation Opportunities IO: An environmental noise d alters some elements of the string ER, and transforms it in a new set of strings called Innovation Opportunities IO. Individual Competent Actors (ICAs) have access only to IO. Clearly, OI is a simple derivative of ER.

Action 6-7: Evaluation and Reward of Proposals made by the Competent Actor CA: The output realized collectively by the ICAs is evaluated by the CE on the basis of an Acceptance Threshold. The reward obtained is distributed among the ICAs that contributed to it, according to the contribution they gave in terms of competences to the successful interpretation. Additional information about this action is provided later.

The three parameters of Length l, Volatility v, and Noise n define the complexity of the competitive environment CE. There is another parameter characterizing the CE, that is the Acceptance Threshold, which measures the intensity of competition, but we discuss this parameter later.