Study On Intelligent Optimization Of Containership Stowage Plan In Full Route

Container shipping has become the best means of transportation in the international economy and trade, it is a main symbol of transport modernization and get fast development worldwide because of its advantages in speediness and security, high quality and low charge. As a key step in container transportation, making stowage plan is to draw an efficient loading and unloading plan for containership that needs the least number of shifts with some constraints. Stowage plan is closely related with the safety and efficiency of shipping. A major objective of shiping business is to make a resonable stowage plan to ensure ship safety, reduce transportation cost, improve efficiency and enhance competition of container transportation.With the rapid increase of the number of containers ship carrying and ports ship calling, containership loading plan optimisation becomes a more and more difficult problem to resolve. It is a multi-objectives combinatorial optimization with complex constraints. Many experts of domestic and foreign contries have done a lot of significant researches in this field and proposed many methods to solve the problem. However, there is few integrated stowage plan model so far. Based on the investigation on quays and the study on many references, the author discussed the characteristic, requirement, process and complexity of containership stowage problem, studied the containership stowage plan in full route, applied the intelligent optimization algorithms to stowage and correlative problems, and proposed some new ideas to solve the problem.Most ships call at more than one port during one voyage. The stowage plan should take into account of the information of containers at each port and the effect of rest calling ports. However, the unknown information of the subsequent ports brings much difficultly to draw stowage plan in full route. A forecast model is proposed to forecast the information of containers of the latter ports, which provides data support for stowage plan in full route. Based on the forecast data, a preplan model in full route of ship calling and bay plan are built. And the general optimization stowage plan is selected using stowage plan fuzzy evaluation model.By analyzing the factors that influence the number of containers ship carrying at each port, a prediction criterion system is raised, Support Vector Machine (SVM) is used to establish nonlinear regression model to forecast the loading and unloading number of containers at each ports. SVM model has the advantages of small samples, high dimensions and strong adaptability andgeneralization ability, which can provide an effective method to multi-effects factors problem such as forecasting the number of containers of subsequent ports.Modelling is the foundation of optimization. Containership stowage problem is so complex that single model cannot satisfy the demands of the issue. To reduce the difficulty, stowage problem is decomposed into two sub problems, i.e., preplan and bays plans. Models of both subproblems are set up respectively and corresponding algorithms are constructed, the final solution is obtainer by combining preplan and bay plan together.Preplan is to groupe containers at current ports to Homogeneous Containers Groups (HCG) according to types, destinations and sizes. Each HCG is assigned to bays on the board of ship based on optimization objectives of preplan. Optimization objectives of preplan are minimizing the number of shifts, maximizing the utilization of cells on ship, and minimizing the number of bays occupied by different destination ports, and some constraints on total longitudinal strength and discrimination of draft etc, which can accomplish even longitudinal distribution of containers on the ship.Bay plan is to assign containers of certain bay to specified cell. According to attribute of destination of containers on the bay, models of both single destination and multi destinations of bay plan are set up respectively. The optimization objectives for bay plan of single destination are minimizing both height of center of gravity and transverse slant moment. Bay plan for single destination is regarded as assignment problem at this stage. Here, taboo search algorithm is used to solve the problem. For the bay with containers of multi ports, optimization objective is minimizing the times of shifts in full route. Implicit graphics heuristic searching algorithm is used to achieve results. The results of simulation prove that the series models can make out an optimal stowage plan suiting actual problem.Containership stowage problem is a complex systematic problem influenced by multitudinous factors of interdependence and inter-restrict. There are lots of stowage plans satisfying criterious constraints at a voyage. To choose the best plan, the criterion system is presented to evaluate the plans, Rough Set theory is used to analyze data, the redundancy decision variables and attributes are removed, and then multi-objectives, multi-hiberarchy fuzzy optimal model is built up to select an general optimization containership stowage plan.Export container operation at yard is closely related with stowage plan. Optimal bay at yard can increase the speed of loading and provide a well foundation for the stowage plan. A model is proposed to load containers at port terminals for the gantry Crane. The model uses the least spanning tree and heuristic method to solve the issue, which can offer theoretic basis for the choosing containers at ecport operation.The final stowage plan is an optimization on the whole. This strategy of drawing stowage plan takes total economic benefit of container shipping as objective and avoids purely pursuing optimization of the current port which might bring the difficulty to the sequel loading and unloading. This model contains three key contents: the forecasts of information of containers, drawing stowage plan and fuzzy evaluating system of containership stowage plan, and the model of containership stowage plan is the most important part among them. Forecast model can forecast the information of sequel containers to provide foundational data for containership stowage plan in full route. Fuzzy evaluation system of containership stowage plans based on rough set theory can be used to selecte the general optimal stowage plan.