Forecasting wind with neural networks

Wind forecasts over a varying period of time are needed for a variety of applications in the coastal and ocean region, like planning of construction and operation-related works as well as prediction of power output from wind turbines located in coastal areas. Such forecasting is currently done by adopting complex atmospheric models or by using statistical time-series analysis. Because occurrence of wind in nature is extremely uncertain no single technique can be entirely satisfactory. This leaves scope for alternative approaches. The present work employs the technique of neural networks in order to forecast daily, weekly as well as monthly wind speeds at two coastal locations in India. Both feed forward as well as recurrent networks are used. They are trained based on past data in an auto-regressive manner using back-propagation and cascade correlation algorithms. A generally satisfactory forecasting as reflected in its higher correlation and lower deviations with actual observations is noted. The neural network forecasting is also found to be more accurate than traditional statistical time-series analysis.     

Conditional modelling of tanker market risk using route specific freight rates

This paper provides statistical evidence in support of the view, widely held in the tanker industry, that there are systematic differences in the degree of risk involved in investing in tankers of different sizes, and in operating tankers in spot and time charter markets. The industry view, broadly supported by the results of this paper, is that larger vessels are 'risker' assets than smaller vessels, and operating vessels in the time-charter market is less risky than employing them on a spot basis. The results are obtained by using a method derived from the financial economics literature, which models both the conditional mean and variance of a variable, known as GARCH modelling. Only one other paper has applied this method to the tanker market, and these results provide confirmatory support of those findings.     

油轮和干散货市场未来运价分析

未来而言，考虑到供求因素，我们认为油轮市场2007年～2008年均将出现同比下滑，但下跌幅度有限。从2009年开始，单壳船的逐渐拆解将支持运价回升，同时，我们预计2007年下半年的油轮运价有望出现季节性反弹；对于干散货市场，基于目前的订单量以及相对较老的运力结构，我们认为2007年～2009年的供求状况将使得运价维持高位，但在2007年，由于一些需求的结构性因素使得BDI指数一路创新高，保守估计，若这些结构性因素在2008年得到缓和，可能会导致2008年运价相比2007年略微回落。     

Wave tranquility studies using neural networks

Information on heights of waves and their distribution around harbor entrances is traditionally obtained from the knowledge of incident wave, seabed and harbor characteristics by using experimental as well as numerical models. This paper presents an alternative to these techniques based on the computational tool of neural networks. Modular networks were developed in order to estimate wave heights in and around a dredged approach channel leading to harbor entrance. The data involved pertained to two harbor locations in India. The training of networks was done using a numerical model, which solved the mild slope equation. Test of the network with several alternative error criteria confirmed capability of the neural network approach to perform the wave tranquility studies. A variety of learning schemes and search routines were employed so as to select the best possible training to the network. Mutual comparison between these showed that the scaled conjugate method was the fastest among all whereas the one step secant scheme was the most memory efficient. The Brent's search and the golden section search routines forming part of the conjugate gradient Fletcher–Reeves update approach of training took the least amount of time to train the network per epoch. Calibration of the neural network with both mean square as well as the sum squared error as performance functions yielded satisfactory results.     

Spillover effects and dynamic correlations between spot and forward tanker freight markets

Monitoring and analysing information transmission across different shipping markets is an important tool for participants to predict shipping freight rates, design portfolio investments and manage freight rate risks. The purpose of this article is to investigate spillover effects and dynamic correlations between shipping spot and derivatives markets (tanker forward freight agreement, FFA) under the multivariate generalized autoregressive conditional heteroscedasticity framework. Empirical results show that spillovers in returns are unilateral from one-month FFA to spot markets, while they are bilateral between one-month and two-month FFA markets. However, insignificant mean spillovers between spot and two-month FFA markets are found. Volatility spillover effects among one-month FFA, two-month FFA and spot freight markets are bilateral. By analysing the correlation between different markets, highly persistent and significantly volatile correlations are found. Moreover, time-varying correlations between one-month and two-month FFA markets are much higher than those of between spot and each FFA market. Results from this article will be helpful to improve participants’ predictions of return, volatility and correlation, which are significant for determining hedge strategies. In addition, the management of freight rate risk and portfolio investment can also benefit from the empirical results obtained in this article.     

Damage detection in offshore structures using neural networks

This paper discusses the damage detection in offshore jacket platforms subjected to random loads using a combined method of random decrement signature and neural networks. The random decrement technique is used to extract the free decay of the structure from its online response while the structure is in service. The free decay and its time derivative are used as input for a neural network. The output of the neural network is used as an index for damage detection. It has been shown that function N is effective in damage detection in the members of an offshore structure. Experimental studies conducted on a reduced model for a real jacket structure with geometrical scale of 1:30 are used. The applied loads were random loads. Two different load spectra were used: White noise, and Pierson-Moskowitz.     

CMAC神经网络在潜艇操纵控制中的仿真应用

基于MATLAB/Simulink软件包.利用拘束模型试验的水动力实验结果.开发了潜艇的六自由度操纵运动预报模型.其典型机动的仿真结果与自航模试验结果吻合较好.在此基础上,成功地将CMAC神经网络技术应用于潜艇操纵运动控制仿真中.仿真结果表明,神经网络控制比单独的PID时的控制效果优越,舵机操舵幅度明显减小,且具有较强的抗干扰能力,鲁棒性强等优点.     

Modeling of ship maneuvering motion using neural networks

In this paper, Neural Networks (NNs) are used in the modeling of ship maneuvering motion. A nonlinear response model and a linear hydrodynamic model of ship maneuvering motion are also investigated. The maneuverability indices and linear non-dimensional hydrodynamic derivatives in the models are identified by using two-layer feed forward NNs. The stability of parametric estimation is confirmed. Then, the ship maneuvering motion is predicted based on the obtained models. A comparison between the predicted results and the model test results demonstrates the validity of the proposed modeling method.     

基于人工神经网络的铁路客运量预测

本文讨论了人工神经网络的特点和建模原理,建立了铁路客运量的预测模型。并用武汉铁路客运量的数据对其进行了验证,得到了较好的预测效果。     

负运价的惩与罚

2007年底,面对中日航运市场日益加剧的无序竞争,交通部再次重拳出击,对安通国际株式会社等四家违规经营公司进行了处罚。从行政罚款到限制航班数量直至取消运输资格,交通部第一次以罚款以外的行政手段整顿和规范中日航线市场秩序。然而,仍然有班轮公司对此表示怀疑,行政手段能否从源头上解决中日航线竞争惨烈的现状？     

Seasonality patterns in the container shipping freight rate market

Shipping indexes have attracted many researchers because they reflect the overall trend of corresponding seaborne markets and can provide implications for the future. Apart from the Baltic Dry Bulk Index (BDI) and correlated indices, the China Containerized Freight Index (CCFI) has been gaining more attention. As a country with large-scale manufacturing, China is an important exporting country and the CCFI was chosen to reflect the container shipping market because the data are more convincing and representative. There have been no systematic attempts to understand the seasonality patterns of container freights. Seasonality patterns reveal the regular fluctuation patterns within a 1-year period. They exist in time series, which are observed more than once a year, like the CCFI. To investigate the nature of seasonality (stochastic and/or deterministic) in container freight rates across different line services, we analyze the CCFI. This paper uses the HEGY method and Monte Carlo method comprehensively to figure out the small sample problem. In addition, seasonal dummy variables are used to test deterministic seasonality. Except for the Japan service series, which contains a half-year unit root, the other container freight rates seem to only involve a non-seasonal unit root at the zero frequency. Deterministic seasonality exists in all the line service series. Furthermore, the seasonality depends on the balance between supply and demand. Under this premise, the seasonal law of freight rates is much obvious.     

Modelling cyclic container freight index using system dynamics

ABSTRACT

This paper investigates the cyclical nature of container shipping market represented by a containerized freight index and proposes a predictive cyclical model of the market. In contrast to the traditional spectral analysis (univariate), system dynamics reflect the drivers of the market in both supply and demand side, and therefore, it is a multi-variate system equilibrium approach consisting of various causal spillovers from sub-components of the market. This study is the first to analyze the cycle of container market using system dynamics. By utilizing system dynamics cyclicality approach, one-step ahead predictions are generated for monthly containerized freight index and compared to conventional benchmarks for post-sample validation. Our study can also help policymakers and shipping liners for better management and invest timing of container ship.     

An analysis of freight rate volatility in dry bulk shipping markets

The world bulk shipping market has been in a peak period since 2003, and this has lasted an incredibly long time considering that the markets are much more complex than before. This paper investigates the characteristics of volatility in dry bulk freight rates of different vessel sizes (capesize, panamax and handysize). The daily returns of freight rate indices of three different types of bulk vessel in the sample period have been examined. The sample period ran from 1 March 1999 to 23 December 2005, and applying the GARCH (generalized auto regressive conditional heteroskedasticity) model showed that the shocks will not decrease but have the tendency to strengthen for all the daily return series. Further, external shocks on the market have a different magnitude of influence on volatility in different types of vessels due to their distinct flexibility. To examine the asymmetric characters of daily return volatility in different bulk shipping sectors and different market conditions, the sample was divided into two periods: one is from 1 March 1999 to 31 December 2002, the other is from 1 January 2003 to 23 December 2005; the EGARCH (exponential generalized auto regressive conditional heteroskedasticity) model was then applied to investigate the asymmetric impact between past innovations and current volatility. The results show that the asymmetric characters are distinct for different vessel size segments and different market conditions. The reasons for the results are discussed and it is considered that the main reasons may be the different flexibility and different commodity transport on different routes. The results from this investigation will be useful for the operators and investors in the dry bulk shipping market to increase profitability and reduce investment risk.     

Modelling forward freight rate dynamics—empirical evidence from time charter rates

The purpose of this paper is to investigate the dynamics of forward freight rate dynamics. We specify our model in a Heath-Jarrow-Morton framework. This model was originally developed for interest rate markets and, in subsequent work, the model has been applied to various commodity markets. We analyse ten years of weekly time charter (TC) rates for a Panamax 65,000 dwt bulk carrier. Our data set consists of 6-, 12- and 36-month TC rates. We use this data to construct, each day, a forward rate function using a smoothing algorithm. We use the smooth data to investigate the factors governing the dynamics of the forward freight rate curve. We find a strange volatility structure in the data. Out results show that the volatility of the forward curve is bumped, with volatility reaching a peak for freight rates with roughly one year to maturity. Also, correlations between different parts of the term structure are in general low and even negative.     

Forecasting the Suez Canal traffic: a neural network analysis

Although the Suez Canal is the most important man-made waterway in the world, rivaled perhaps only by the Panama Canal, little research has been done into forecasting its traffic flows. This paper uses both univariate ARIMA (Autoregressive Integrated Moving Average) and Neural network models to forecast the maritime traffic flows in the Suez Canal which are expressed in tons. One of the important strengths of the ARIMA modelling approach is the ability to go beyond the basic univariate model by considering interventions, calendar variations, outliers, or other real aspects of typically observed time series. On the other hand, neural nets have received a great deal of attention over the past few years. They are being used in the areas of prediction and classification, areas where regression models and other related statistical techniques have traditionally been used. The models obtained in this paper provide useful insight into the behaviour of maritime traffic flows since the reopening of the Canal in 1975—following an 8-year closure during the Arab-Israeli wars (1967-1973)—till 1998. The paper also compares the performance of ARIMA models with that of neural networks on an example of a large monthly dataset.     

Forecasting the Suez Canal traffic: a neural network analysis

Although the Suez Canal is the most important man-made waterway in the world, rivaled perhaps only by the Panama Canal, little research has been done into forecasting its traffic flows. This paper uses both univariate ARIMA (Autoregressive Integrated Moving Average) and Neural network models to forecast the maritime traffic flows in the Suez Canal which are expressed in tons. One of the important strengths of the ARIMA modelling approach is the ability to go beyond the basic univariate model by considering interventions, calendar variations, outliers, or other real aspects of typically observed time series. On the other hand, neural nets have received a great deal of attention over the past few years. They are being used in the areas of prediction and classification, areas where regression models and other related statistical techniques have traditionally been used. The models obtained in this paper provide useful insight into the behaviour of maritime traffic flows since the reopening of the Canal in 1975—following an 8-year closure during the Arab–Israeli wars (1967–1973)—till 1998. The paper also compares the performance of ARIMA models with that of neural networks on an example of a large monthly dataset.     

Marine accident analysis for collision and grounding in oil tanker using FTA method

In this study, collision and grounding data registered in GISIS (Global Integrated Shipping Information System) were investigated for oil tankers. The database includes the information of the collision and grounding accidents during the period between 1998 and 2010 in oil tankers. The risk assessments were carried out using fault tree analysis (FTA) programme for the incidents as collision and grounding occurred in oil tankers. In this study, we were able to investigate first the potential problems which cause the collision and grounding accidents have been determined, second, the occurrence of accidents has been shown with causal factors by the FTA method, and, finally, the significance degree of the initial events causing occurrence of accidents have been put forth. Collision in oil tanker resulted in economical loss (81%), pollution (6%) and death or injury (13%). Grounding in oil tanker resulted in economical loss (91%) and pollution (9%). According to the FTA results, the main reason for the accidents originating from human error is as follows: for collision accidents, Convention on the International Regulations for Preventing Collisions at Sea (COLREG) violation and the lack of communication between vessels; and for grounding accidents, the interpretation failure of the officer on watch and lack of communication in the bridge resource management.     

基于人工神经网络的舰船稳定性分析

船舶在海上经常受到风浪流雾等不可估测因素的影响,控制船舶的稳定性非常重要。本文通过分析单个人工神经元PID控制和多个神经元PID控制方式设计基于人工神经网络PID控制算法,并将此算法应用于船舶稳定性控制。仿真实验说明,人工神经网络应用于船舶稳定性控制的有效性和准确性。     

Simulation of the mean zero-up-crossing wave period using artificial neural networks trained with a simulated annealing algorithm

The aim of this work was to develop a predictive model to forecast the mean zero-up-crossing wave periods (T _z ) for 3-hourly sea states at a location in the Pacific using artificial neural networks (ANNs). Seven multilayer ANNs were trained with a simulated annealing algorithm. The output of each trained ANN was used to estimate each of the seven parameters of a new distribution called the hepta-parameter spline proposed for the conditional distribution of T _z , given some mean zero-up-crossing wave periods and significant wave heights. After estimating the parameters of the distribution, the model was used to simulate and predict future values of T _z . Forecasting a sea state and developing the joint distribution of sea state characteristics with the help of the simulated characteristics are also discussed in this article.     

A new approach for deriving temperature and salinity fields in the Indian Ocean using artificial neural networks

This work reports a new methodology for deriving monthly averages of temperature (T) and salinity (S) fields for the Indian Ocean based on the use of an artificial neural network (ANN). Investigation and analysis were performed for this region with two distinct datasets: (1) monthly climatological data for T and S fields (in 1° × 1° grid boxes) at standard depth levels of the World Ocean Atlas 1994 (WOA94), and; (2) heterogeneous randomly distributed in situ ARGO, ocean station data (OSD) and profiling (PFL) floats. A further numerical experiment was conducted with these two distinct datasets to train the neural network model. Nonlinear regression mapping utilizing a multilayer perceptron (MLP) is employed to tackle nonlinearity in the data. This study reveals that a feed-forward type of network with a resilient backpropagation algorithm is best suited for deriving T and S fields; this is demonstrated by independently using WOA94 and in situ data, which thus tests the robustness of the ANN model. The suppleness of the T and S fields derived from the ANN model provides the freedom to generate a new grid at any desired level with a high degree of accuracy. Comprehensive training, testing and validation exercises were performed to demonstrate the robustness of the model and the consistency of the derived fields. The study points out that the parameters derived from the ANN model using scattered, inhomogeneous in situ data show very good agreement with state-of-the-art WOA climatological data. Using this approach, improvements in ocean climatology can be expected to occur in a synergistic manner with in situ observations. Our investigation of the Indian Ocean reveals that this approach can be extended to model global oceans.