Simulation of aquatic food web and species interactions by adaptive agents embodied with evolutionary computation: a conceptual framework

Individual-based and state variable-based adaptive agents (AA) are discussed regarding their relevance to different types of ecosystems. Individual-based AA proved applicable to a spatially explicit simulation of highly simplified terrestrial food webs. State variable-based AA with evolutionary computation (EC) embodied are suggested for the simulation of aquatic food webs and plankton species interactions. Embodiment of EC in AA can be achieved by evolving predictive rules (ER), differential equations (EDE) or artificial neural networks (ANN) derived from a diverse lake database. In order to provide ecosystem simulation with resilience to environmental change, agent banks can be created containing alternative agents for same species or functional groups from different lakes. State variable-based AA are currently tested for aquatic ecosytem simulation by means of a diverse lake database. It promises to overcome constraints by the rigidity of traditional lake ecosystem models.