Page 71 - Getting the Picture Modeling and Simulation in Secondary Computer Science Education

Defining and Observing Modeling and Simulation in Computer Science
decided what details to highlight and what details to ignore. Problem is formulated in a way that enables us to use a computer and other tools to help solve them (CSTA Computational Thinking Task Force, 2011).
2. Formalization stage: a computer program is constructed, i.e.
requirements and specifications are stated and the system is
implemented and tested (Comer et al., 1989). This includes making
pilot runs, verifying the program and checking validity of the
simulation model. If necessary the program is adjusted (Law, 2015). 3 Thus, the formalization stage is a cyclic process in itself.
3. Execution stage: the model is being used for its purpose: designing and running experiments (Law, 2015).
Simulation modeling encompasses three methods: (1) System dynamics, associated with high level of abstraction where the individual objects are aggregated. The models can be described in terms of differential equations that are often non-trivial to solve. (2) Discrete event modeling, where the system modeled is considered to be a process, “i.e. a sequence of operations being performed across entities”. The level of abstraction is lower. (3) Agent based modeling (ABM), which is made possible with recent growth of availability of CPU power and memory, does not assume any particular abstraction level. Agents have their properties and behavior and one can start building a model by identifying agents and describing their behavior even without knowing how a system behaves as a whole. ABM makes it possible to model systems that are difficult to capture with older modeling approaches (Borshchev, 2013) and it does not require familiarity with differential equations or mathematics beyond reach of secondary students. In our view, the characteristics of the ABM make it a suitable modeling method for our students who often lack deep understanding of the phenomena they model and make models specifically to deepen their understanding. To conclude, we consider conceptual representation which could be realized through the employment of ABM methods and software, in which “you give computational rules to individual agents and then observe, explore analyze the resultant aggregate patterns” (Wilensky, 2014) suitable for use in a secondary CS class “because the individual- level behavior of agents is relatively simple, [and] ABMs feature relatively simple computer programs that control the behaviors of their computational agents” (Wilensky, 2014).
69