The behavior of complex ecological systems is often analyzed by means of numerical simulation. Traditionally, system components are populations and concentrations, whose quantitative dynamics are described by differential or difference equations. These model formalisms cannot account for the variability typical of ecological systems. In contrast, variable-structure modeling can be performed with simulation tools allowing for dynamically changing numbers of interacting components. Although populations may constitute these components, modern computational tools make it feasible to depart from the lumped population level, and describe system behavior in terms of interacting individuals. Individuals, as unique, discrete and local entities, have the additional advantage of capturing much of the ecological variability. In this paper, individual-oriented simulation is discussed in relation to a discrete-event simulation tool developed in object-oriented Smalltalk. In discrete-event simulation the autonomous character of individual behavior comes out well. Objects in object-oriented programming are conceptually close to individuals. An application is described based on the familiar Nicholson-Bailey equations for a host-parasitoid system. It is shown how, with a minimum in programming effort, an individual-oriented model of this system is derived from a basic ecological simulation framework. Population dynamics emerging from host and parasitoid individual interactions, in a patchy environment, are analyzed for some elementary assumptions and parameter values. Parasitoid efficiency in locating and exterminating local host populations, together with system size in terms of the number patches, appear major factors determining the possibility of prolonged persistence of the metapopulation. The perspectives of the object-oriented approach are discussed. It is concluded that interactive, object-oriented programming systems constitute a valuable extension of the ecological modeler's toolbox.
