Simulation and object-oriented programming: the development of SERB

Abstract

An object-oriented programming language (OOPL), Flavors, was used to implement a modeling language and ancillary simula tion tools designed specifically to meet the needs of research biologists. A biological system is typically considered as a set of interacting entities, an abstraction which closely matches the pro gramming paradigm of Flavors. Some of the important charac teristics of Flavor programming are presented. Then SERB (the Simulation Environment for Research Biologists) is described. SERB uses Flavor objects to represent biological entities, resulting in a high degree of transparency between model and simuland. The simulationist creates model components from a library of component types and defines a model as some interacting set of these components. These models can then be used in simula tions or by higher level functions, such as sensitivity analysis. Flavors was found to be an ideal language for implementing modeling and simulation software for this domain.

Objects for Simulation: Smalltalk and Ecology*

“…Already in SIMULA, new models could be constructed efficiently, by defining them as subclasses of existing model classes. Object-oriented programming has been applied successfully to deal with the complexity of ecological systems (Larkin et al, 1988;Saarenma et al, 1988).…”

Simulation of insect disease dynamics: an application of SERB to a rangeland ecosystem

“…One of the primary factors limiting the use of these quantitative techniques by biologists is the lack of training in the area of mathematics and computer science. As discussed by Larkin et al (1988) the object-oriented and highly interactive simulation environment developed in SERB aids non-mathematical and non-computer-oriented biolo- gists in constructing and using complex computer models to synthesize and evaluate their research problems.…”

“…In the E. grylli portion of the model individual stages of the pathogen are identified and serially arranged to represent maturation from one life stage to the next. Interaction of the two organisms is modeled similarly with specialized components (Carruthers et al 1986;Larkin et al 1988) used to simulate host and pathogen development simultaneously. Flow from one stage to the next in both organisms is regulated by environmental conditions.…”

“…those involving multiple simulation runs), such as sensitivity analysis, multi-season evaluations, optimization, etc. are relatively trivial programming problems in SERB as the model itself is implemented as an instance of a simulation object (a defmodel) and thus can be manipulated using the same message and methods-handling techniques as invoked in lesser system components (see Larkin et al 1988). …”

“…We feel systems analysis and simulation are extremely useful methodologies that can not only help provide insight into the structure and operation of complex biological systems but also help with overall system synthesis and mangement. Here we present a simulation study, developed in SERB (Larkin et al 1988), of a naturally occurring host/pathogen life system and discuss analyses that have been useful in understanding the dynamics of this disease. Specific attention will be given to the transition from conceptualization of the biological simuland to actual model formation and use, accentuating features in SERB that facilitate this process.…”