Comparison of MATLAB, Simulink and AnyLogic Approach to ARGESIM Benchmark C9 ‘Fuzzy Control of a Two Tank System’

Simulation Notes Europe SNE 23(3-4), 2013, 195-200
DOI: 10.11128/sne.23.bn09.10219


This contribution compares modelling and simulation of the ARGESIM Benchmark C9 ‘Fuzzy Control of a Two-Tank System’ with three approaches: (1) programming directly in MATLAB (2) using SIMULINK and the MATLAB Fuzzy Toolbox, and (3) using AnyLogic, a Java-based grapic simulation environment.  The MATLAB implementation required direct program-ming, whereby the nonlinear ODE model for the two-tank system was simulated by MATLABs ODE solvers, and fuzzification, inference, and defuzzification was pro-grammed by ‘pure’ vector handling feature. The Simulink implementation is straightforward: graphical blocks for the ODE model, and use of the Fuzzy Toolbox, wich sup-ports graphical design of the fuzzy controller. Anylogic offers various graphical modelling methods, also classic block diagrams. But for tis comparison the System Dynamics modelling capability was used, which allows a genuine mapping of ‘tanks’ as reservooir variables; the fuzzy controller was programmed directly in Java em-bedded into the simulation environment. The contribution discusses advantages and disadvantages of the modelling approaches – in modelling, in implementation and in simulation and efficiency.