A generic model of hierarchy for systems analysis and simulation

Problems of global change and sustainable agriculture have motivated systems analyses that span increasing numbers of hierarchical levels. The reuse of existing simulation models in these expanded analyses is appealing, though experience with crop models shows that the design of the models call limit their reuse. In order to avoid these design problems, a new framework for systems analysis was developed based on hierarchy theory. The framework formalizes a metaphor used to define hierarchy: each hierarchical object (holon) is analyzed and modeled with two faces, named up and down, through which the object can communicate data; up is a portal for data exchange with the next higher level and down provides data exchange with the object's subsystems, which are also holons. This generic model of hierarchy was coded in object-oriented computer languages and was developed into a programming framework named JanuSys. JanuSys organizes communications between modeled holons and provides an orderly sequence to their simulation. The framework permits dynamic creation of a user-defined number of hierarchical levels, containing user-defined numbers of subsystems in each holon. These capabilities, combined with the generic model of hierarchy, are required for testing basic theories of hierarchy and systems evolution. If hierarchy theories are correct, JanuSys will provide a more successful foundation for systems analysis than non-hierarchic approaches. (C) 1998 Elsevier Science Ltd. All rights reserved.