The model builder
ILCYM (version 2.0) software provides a “model builder” for facilitating the determination of the best-fitting functions to achieving high accuracy of the model and for the compilation of the final model (model design). Further, it provides tools for validating models by comparing simulation results with field-derived life-table data. It should be noticed that the model builder might currently not provide solutions for each insect species of interest or under study. However, follow-up versions will be improved to adjust to a wider range of pest species of different insect orders and families.
The “model builder” interactively facilitates the description of temperature-driven processes in insect development and the selection of predefined, best-fitting functions. These functions are model components, henceforth called “sub-models” of the overall generic phenology model, or “modules” (the term “modules” is used because a sub-model can be replaced by another function that might describe better the specific process in the system), which are automatically implemented in the overall pest phenology model. It provides statistical outputs that help users to choose appropriate sub-models and interpret and validate models and simulation outputs. Consequently, it enables non-experts to use advanced modeling techniques in their research. The ILCYM “model builder” delivers some kind of standardization in modeling insect pest populations. This has several major advantages. It reduces programming efforts and provides a reasonable framework for pest phenology modeling. It also offers the possibility of reusing the model design, i.e. using the framework redefining the parameter only. However there are also disadvantages, e.g. restricting the modeler somehow to certain modeling approaches, reducing scientific freedom and not providing solutions for each problem. It is faster to produce a pest phenology model with ILCYM, but less flexible and it might be difficult to include further model components that are not predefined in the ILCYM program. Furthermore it should be noted that ILCYM is not a good visual modeling environment. It is restricted to the standard model design (approach) implemented in the program. If users want or need to change the design of the overall model, other modeling programs are available which are specifically created for that purpose (i.e. Simile, Stella, ModelMaker and PowerSim). Simile can be recommended; it provides good model structure visualization and is very flexible for model modifications and extensions. However, these programs do not have the analytic and statistical tools for selecting appropriate models by fitting sub-functions to experimental data. Here, ILCYM is quite strong and provides users with summary output tables showing statistical analysis and high resolution graphics.
The ILCYM “model builder” focuses strongly on “parameter determination” for models. The program produces outputs (statistics and graphs) that help to find the best parameters for a specific species on a statistical basis. Parameters can be stored in a database so that a model can be concluded step-by-step until its finalization, and finally the model can be evaluated (using validation and sensitivity analysis of the model) and modified by using several tools provided in ILCYM. Once the parameters are defined, they can be used in other modeling programs (as mentioned above).