Object-oriented simulation of integrated whole farms: GPFARM framework

Simulation frameworks for Decision Support Systems (DSSs) at the whole-farm level have not adequately supported process level management and integration of complex farming systems. Development of the Great Plains Framework for Agricultural Resource Management (GPFARM) DSS for whole-farm management required a simulation package that could handle this complex system. Object-oriented (OO) techniques now offer improved opportunity for the development of suitable whole-farm simulations. A whole-farm simulation framework was developed using Object-oriented programming (OOP) and executes appropriate simulation modules, written in procedural languages (FORTRAN and BASIC), for the lower-level processes. Abstraction, encapsulation, and hierarchy were crucial to simplifying whole-farm complexity. OOP relationships, particularly inheritance in crops, animals, and events were key in allowing dynamic (runtime) setup and simulation of the farm system. Event objects were positioned outside of the simulation class and farm state to provide event input format flexibility and allow the framework to check the system state before events were implemented. The inclusion of existing or extended procedural modules cut development time, helped insure maintenance support from cooperators, and assisted with deployment of computationally efficient code. The resulting package simulates whole-farm management involving interactive land units, integrated crop and livestock operations, crop rotation systems and multiple commodities, and soil and climate variability. Simulation of whole-farm scenarios provides information for DSS applications to display and compare management alternatives. This research showed that an effective OO framework for integrated farming systems should consist of a thorough object model of a farm state, a flexible input and implementation of events, and a simulation environment to accomplish the biological, chemical, and physical simulation of a farm.