Research On Bulk Port Stack-Scheduling System Based On Ontology And Evolutionary Algorithm

With the development of economic integration and the trend of globalization, rapid growth has been seen in our economy, foreign trade as well as the volume of bulk freight. The increasing requirements of bulk freight and the diversity of clients’ demands not only bring opportunities to port enterprises, but also pose great challenges to the port services. Therefore, it becomes critical to further development of bulk port that how to use the existing port facilities by the adoption of modern logistics management mode to enhance the construction and application of port information to ease the pressure of port throughput, to improve the efficiency of operations of loading and uploading and service of bulk port.Stack-Scheduling is one of the core businesses of port enterprises, especially for the southern bulk port of China. Due to the extremely scarce resources of stack and diversity stockpiling site of different cargo, proper distribution of stacks could improve port utilization, increase the efficiency of overall operations of port and waterway logistics, and lower the cost of operations to better satisfy the clients. Based on the scheduling complexity of southern bulk port of China with multi objectives and restrictions, and with references to The Operation Management System and General Software Industry of Guangzhou Port Group (2008B090500244), The Application Trial Project Based on Motor Ro-Ro Management System of RFID Port (2009B090300467), and The Optimization Research of Integration and Scheduling of Logistics Resources (71132008), which is a key program of National Natural Science Foundation, this dissertation conducted a deep analysis of the features and status quo of stack-scheduling of southern bulk port of China by a comprehensive use of theories and methodologies of ontology, ontology reasoning, BP-ANN and genetic algorithm, designed and built a smart scheduling system of bulk port stack based on ontology and evolutionary algorithm. And the main research contents and results are as follo wings:(1) Framework of scheduling system of bulk port based on ontology and evolutionary algorithmBased on operating rules of stack-scheduling of southern bulk port of China and optimization of objectives, and considering the problems of stack-scheduling over dependent on workers’experiences, lack of optimization standards, planning and feedback mechanism, this dissertation initiated Ontology and Evolutionary Algorithm Based Bulk-Port Stack-Scheduling System (OEABSS) with the method of qualitative reasoning and quantitative calculation to solve the problem of multi-objective optimization scheduling of bulk port under the condition of many restrictions and influencing factors. Ontology was used in this system to achieve restriction reasoning of ontology-based cargo and stack. BP-ANN and genetic algorithm was applied for the settlement of complicated problems of traffic forecasting and stack distribution.(2) Mode of forecasting the traffic of bulk port based on multi-improved BP-ANN.Based on the analysis of the existing problems and improvements of BP-ANN, in accordance with the features of traffic forecasting of bulk port, and from the aspects of sample selection, pre-treatment, the confirmation of BP structure, the initialization of weights and thresholds, and network training simulation, etc. this dissertation initiated Multi-Improved BP-ANN Forecast Model Building (MBPFB). With the application of various optimization methods and improvements including compensation strategy, comparison, genetic algorithm and sliding windows, this dissertation better resolved the over-fitting problems easily coming up in the process of forecasting, enhanced the capacity of nonlinear forecasting of handling uncertainty in ANN processing and sample fluctuation, which has a wild generalization.(3)Intersect domain ontology building and integration methodologyBased on the existing ontology building and integration methodology and taking the features of intersect domain ontology in to consideration, this dissertation initiated4principles of intersect domain ontology building and Intersect Domain Ontology Building from above to below. For intersect domain ontology integration, this dissertation initiated cross domain ontology integration system structure and Intersect Domain Ontology Integration Method. This ontology building and integration methodology served to improve the efficiency of cross domain ontology building, lower the complexity of building, solve the re-use difficulty of ontology, and largely realize the sharing of ontology.(4) Method for solving problems of complicated multi-objective optimization based on improving NSGAIIBased on the research of methods for solving problems of multi-objective optimization and in accordance with the huge searching space, restrictions and influencing factors of bulk port, this dissertation initiated the multi-objective optimization method based on improved NSGAII. Through the application of improved methods of restrictions on matrix, random repaired operators, adaptive crossover based on genetic algebra and mutation rate, the capacity of multi-objective optimization was enhanced under the condition of many restrictions and influencing factors to avoid algorithm from local optima, which increased algorithm convergence efficiency.(5) Development of OEABSS prototype systemBased on former research, this dissertation achieved the building of OEABSS prototype system and realized the core functions of basic date management, stack matching evaluation and smart stack-scheduling by applying joint technology framework of J2EE-MVC, Hibernate and Spring and integrating key technology of RCP, Protege and Jess.