|
Shipping market behavior and decision making patterns illustrated through 3-D graphs, or "decision surfaces". This one plots a tanker freight rate as a continuous function of the tanker supply and demand variables (Parsons and Li 1997). |
Marine systems management problems are formulated and solved using a variety of quantitative techniques, such as deterministic and probabilistic optimization and/or decision analysis. The correct formulation and solution of a wide range of optimal fleet deployment problems for different types of commercial marine transportation (dry bulk carriers, tankers, and containerships) is one example. Another is the minimal cost of time routing and scheduling of ships under known or probabilistic weather conditions with prohibited sailing regions and multiple destinations. Port and inland waterway planning and operations problems have been studied from two viewpoints. For example, statistical determinations of key variables as a function of deadweight to provide models for port planning executive decisions or from policy analysis such as the investigation of the effect of technical and management improvements in the recent reduction of inland waterway fleet fuel consumption and the corresponding decrease in fuel tax revenues, despite the marked observed increase in ton-miles of cargo carried.
Artificial intelligence and expert systems have been applied to ship operations problems such as the optimal container stowage problem. Detailed management information systems have been developed for the same problem as well as for reliability databases in use by the Great Lakes Shipping industry, or, more recently, by the United States Coast Guard. Recent research also includes a comprehensive study of human error as a major factor in marine accidents, in fire prevention and response for cruise ships. Topics studying marine engineering problems using marine systems methodology currently include the evolution of condition-based maintenance policies as opposed to following manufacturers' recommendations for standardized marine engines, and integrated cost/benefit analysis and economic studies involving environmental considerations. The tanker industry in the US and abroad, OPA 90, IMO regulations, ISM, etc., are also of interest. Fuzzy systems and neural networks have also been utilized to improve shipbuilding market forecasting. In an integrated market forecasting system, neural networks are used to forecast the level of economic variables 24 months into the future. Fuzzy decision modelers are then used to replicate aggregate ship purchase, second-hand sales, and scrapping market decisions based upon these conditions. Genetic Algorithms (GAs) have been used successfully to design membership functions and decision rules for the fuzzy decision modelers which will reproduce historical data in a least squares sense.
|
|
Shipping market behavior and decision making patterns illustrated through 3-D graphs, or "decision surfaces". This one plots a tanker freight rate as a continuous function of the tanker supply and demand variables (Parsons and Li 1997). |
FACULTY: Parsons, Perakis