The contribution focuses on the design of a control algorithm aimed at the operative control of runoff water from a reservoir during flood situations. Management is based on the stochastically specified forecast of water inflow into the reservoir. From a mathematical perspective, the solved task presents the control of a dynamic system whose predicted hydrological input (water inflow) is characterised by significant uncertainty. The algorithm uses a combination of simulation model data, in which the position of the bottom outlets is sought via nonlinear optimisation methods, and artificial intelligence methods (adaptation and fuzzy model). The task is written in the technical computing language MATLAB using the Fuzzy Logic Toolbox.
This paper compares Fuzzy Inference System (FIS), Support Vector Machine (SVM) and MultiLayer Feed-forward neural network (MLF) in modeling a driver's decision when making a discretionary lane changing move on a freeway. The FIS model has been developed and published in an earlier work by the authors, whereas the SVM and MLF models are newly developed in this research. The FIS, SVM and MLF models use the same four inputs: the gap between the subject vehicle and the leading vehicle in the original lane, the gap between the subject vehicle and the leading vehicle in the destination lane, the gap between the subject vehicle and the trailing vehicle in the destination lane, and the distance between the preceding and trailing vehicles in the destination lane. The models give a binary decision of "no, stay in the same lane" or "yes, move to the destination lane now". These models were trained and then tested with the Next Generation SIMulation (NGSIM) vehicle trajectory data. The results have shown that the FIS has the highest accuracies in making correct lane changing decisions. It recommends "yes, move to the destination lane now" with 82.2% accuracy, and "no, stay in the same lane" with 99.5% accuracy. The SVM model also outperformed the traditional gap acceptance model which was used as the benchmark. However, the MLF model was not as accurate as the gap acceptance model.