| The problem of shipping emergency command, decision and scheduling is a hot topic in the field of maritime research. It falls into the category of system science in the view of control. In this thesis, the various ship devices, human activities and events, which interact with each other, are regarded as a system. Using Petri nets as a tool, the system is modeled and analyzed in order to improve the method of handling maritime accident efficiently and reduce the accident caused by the mistakes in human decision-making.At first, the basic of Petri nets theory and the concept of emergency control system for shipping (ECSS) are introduced. ECSS is then modeled by using generalized stochastic Petri nets (GSPN), a subclass of Petri nets, and its open-loop system performances are obtained. With the aim of solving deadlock and resource competition, the closed-loop system is finally obtained by introducing inhibitor arc, finite capacity place and so on in the open-loop system. The PIPE based simulation results show that the closed-loop system is able to well model the real ECSS and offer suggestions of emergency command decision-making for the maritime sector.Secondly, the reasons of human decision-making mistakes from the view of error chain are analyzed in details based on the got results of studying ECSS model. For the first time, the improved TOPSIS (Technique for Order Performance by Similarity to Ideal Solution) algorithm which integrates the "integrated weight" and "orthographic projection" is proposed in this thesis and applied to the ship emergency decision-making model. The result of simulating a real ship collision accident indicates that the improved TOPSIS algorithm is able to avoid human error for decision-making and improve the real-time performance of the computer-aided decision-making system, and realize the original design objectives of the ECSS.Finally, the conclusion remarks are presented and the future work is discussed in the end of the thesis. |
PDF Full Text Request