FLOOD STAGE FORECASTING WITH SUPPORT VECTOR MACHINES1

ABSTRACT: Machine learning techniques are finding more and more applications in the field of forecasting. A novel regression technique, called Support Vector Machine (SVM), based on the statistical learning theory is explored in this study. SVM is based on the principle of Structural Risk Minimization as opposed to the principle of Empirical Risk Minimization espoused by conventional regression techniques. The flood data at Dhaka, Bangladesh, are used in this study to demonstrate the forecasting capabilities of SVM. The result is compared with that of Artificial Neural Network (ANN) based model for one‐lead day to seven‐lead day forecasting. The improvements in maximum predicted water level errors by SVM over ANN for four‐lead day to seven‐lead day are 9.6 cm, 22.6 cm, 4.9 cm and 15.7 cm, respectively. The result shows that the prediction accuracy of SVM is at least as good as and in some cases (particularly at higher lead days) actually better than that of ANN, yet it offers advantages over many of the limitations of ANN, for example in arriving at ANN's optimal network architecture and choosing useful training set. Thus, SVM appears to be a very promising prediction tool.     

Demographic factors and land-use planning in the small islands of Southern Europe

Since the end of the 1970s, the southern European countries have shown an exceptional reduction in fertility rate. From the highest levels among the developed nations, these countries dropped beneath the substitution rate level: in Greece there is an average of about 1.5 children per woman, and Italy (starting three to four years ago), with 1.3 children per woman, is now the country with the lowest fecundity rate in the world. Land-use planning in southern European small islands therefore requires substantial revision. In the areas where western civilization began, which are highly populated and have a long history, cultural and ethnic aspects of tradition are fundamental to environmental management and to the defense of historical heritage. They also place a strong value on sustaining tourism, the most relevant economic activity, that allows them to survive and maintain a high welfare level. For some decades they have had populations with a marked presence of young people and high emigration rates, but now they are fast becoming dominated by the elderly and must prepare for a period of fast reduction in youth of the workforce, while the peripheral areas of Asia and Africa are entering a sudden demographic growth phase. The demographic structure has also been deeply altered both by previous migrations and by random variations, as usually happens in all small communities. Social services for younger and older people have had to be adapted rapidly, reorganizing high-school management, hospital and health-care structures, in-house assistance, and so on. There is a need to rethink the job market and favor the immigration of highly specialized workers, which is a necessity for technical evolution. Sustainable development is constrained nowadays not only by the scarcity of natural resources, but also by the quality and quantity of human resources. Proper policies for population and land-use planning are highly correlated factors; they have to be considered with respect of these new, rapid demographic changes.     

A REGIONAL APPROACH TO PROJECT EVALUATION1

ABSTRACT. Water and related land resource planning has generally been characterized by project evaluation in isolation. Feasible alternatives have been ignored because the economic analysis did not include the interdependence of such subregional and regional variables as factor and product prices and production possibilities. This paper presents an economic framework, consistent with a regional development objective and an efficiency criterion, for proposing and evaluating resource projects. Subregional and regional derived demand curves for water are developed under alternative assumptions of subregional competition, regional market restraints, and yield and price uncertainty. The derivation of water demand curves using the proposed regional framework as compared to sub regional isolation has the desirable properties of (1) a marginal analysis that is more flexible over time is substituted for a static average analysis; (2) range estimates incorporating probabilities are substituted for point estimates; (3) approximations to functional demand curves are substituted for “needs”; and (4) fewer resources are required to meet a regional market restraint.     

NEW STRATEGIES FOR AGRICULTURAL DEVELOPMENT IN CHINA''''S MOUNTAIN AREAS

Ｂｙｔｈｅｅｎｄｏｆｔｈｅ 2 0ｔｈｃｅｎｔｕｒｙ ,ｔｈｅｍａｊｏｒｉｔｙｏｆｍｏｕｎｔａｉｎｒｅｇｉｏｎｓｉｎＣｈｉｎａ ,ｅｓｐｅｃｉａｌｌｙｔｈｏｓｅｐｏｖｅｒｔｙ ｓｔｒｉｃｋｅｎａｒｅａｓ,ｈａｄｓｏｌｖｅｄｔｈｅｐｒｏｂｌｅｍｓｏｆｆｏｏｄａｎｄｃｌｏｔｈｉｎｇ .Ｔｈｅｙａｒｅｎｏｗｍａｒｃｈｉｎｇｔｏｗａｒｄｔｈｅｇｏａｌｏｆｌｉｖｉｎｇａｗｅｌｌ ｏｆｆｌｉｆｅ .Ｎｅｗｃｈａｌｌｅｎｇｅｓｔｈａｔｔｈｅｍｏｕｎｔａｉｎｒｅｇｉｏｎｓａｒｅｆａｃｅｄｗｉｔｈｉｎｔｈｅ 2 1ｓｔｃｅｎｔｕｒ…     

General prevention and risk minimization in LCA A combined approach

Methods for life cycle assessment of products (LCA) are most often based on the general prevention principle, as opposed to the risk minimization principle. Here, the desirability and feasibility of a combined approach are discussed, along with the conditions for elaboration in the framework of LCA methodology, and the consequences for LCA practice. A combined approach provides a separate assessment of above and below threshold pollution, offering the possibility to combat above threshold impacts with priority. Spatial differentiation in fate, exposure, and effect modelling is identified to play a central role in the implementation. The collection of region-specific data turns out to be the most elaborate requirement for the implementation in both methodology and practice. A methodological framework for the construction of characterization factors is provided. Along with spatial differentiation of existing parameters, two newly introduced spatial parameters play a key role: the sensitivity factor and the threshold factor. The practicability of the proposed procedure is illustrated by an example of its application. Providing a reasonable data availability, the development of separate LCA characterization factors for the respective assessment of pollution levels above and below environmental threshold values seems to be a feasible task that may add to LCA credibility.     

发电设备可靠性研究现状与发展

通过对国内外发电设备可靠性研究的分析,提出了我国发电设备可靠性研究大体经历了可靠性指标统计分析、传统可靠性模型分析以及电力市场下可靠性分析3个阶段。在比较各阶段研究特点的基础上,指出了建立发电设备可靠性成本与可靠性效益之间的平衡,设定反映二者关系的新的可靠性指标是电力市场下发电设备可靠性研究的新特点。笔者认为,所提出了数据挖掘技术,是发电设备可靠性管理实现实时化、智能化与预测化的一种有效方法。     

商贸市场火灾风险评估方法

通过对典型商贸市场的实地调研,依据历年商贸市场火灾事故统计数据,采用将统计方法与专家评分相结合的方法,运用系统安全及模糊数学的相关理论,提出了商贸市场火灾风险评估方法,从而为商贸市场的安全管理及火灾风险评价提供了切实可行的参考依据。     

Water Network Synthesis Using Mutation-Enhanced Particle Swarm Optimization

Different techniques for the synthesis of industrial water reuse/recycle networks have been developed in recent process integration research. These tools range from graphical pinch analysis approaches to mathematical programming models. The latter have the advantage of being flexible enough to incorporate various water network constraints, but in many cases these are often non-linear, thus making the identification of global optima difficult. Recent work has demonstrated the effectiveness of metaheuristic algorithms such as particle swarm optimization (PSO), for finding good solutions these problems. This work describes the use of a modified PSO for solving mixed integer non-linear programming (MINLP) models for water network synthesis. By incorporating a mutation operator for the binary variables in the model, the algorithm is able to escape sub-optimal network topologies and proceed towards better solutions than can be found with ordinary PSO. Two case studies involving water recycle/reuse are used to demonstrate the new design methodology.     

THE IMPACT OF HIGH INTEREST RATES ON OPTIMUM MULTIPLE OBJECTIVE DESIGN OF SURFACE RUNOFF URBAN DRAINAGE SYSTEMS1

ABSTRACT. This research examines the impact of high rates of interest upon the least cost system design for urban drainage systems when water quality is a critical parameter. Results of examination of twelve alternative system designs in a case study watershed indicate the least cost study design is highly sensitive to the rate of interest, but not sensitive to the water quality parameters. When the high rates of interest currently prevalent are introduced into the model those systems which contain open channel collection components are selected as the least cost system. At low rates of interest pipeline collection components are selected as the least cost system. Holding pond components of system design are cost effective at several levels of water quality. They are neutral to the rate of interest so they are incorporated in least cost systems at all the levels of interest rates. The results of the study indicate that at the current high rates of interest open channel collection systems and holding ponds are cost effective system components to achieve selected levels of water quality in urban drainage system design.     

IMPROVING IRRIGATION RETURN FLOW QUALITY WITH A WATER RENTAL MARKET1

ABSTRACT: Current policies for correcting the problem of irrigation return flow pollution tend to attack the symptoms of the problem, rather than its cause. The present institutional arrangement for allocating irrigation water is seen as the source of the problem. This paper examines the water quality benefits of altering the institutional arrangement to allow for irrigation water transfers through a rental market. It is conceptualized that by creating a water rental market an opportunity cost would be associated with the use of irrigation water such that profit maximizing farmers would be induced to use his water supply more efficiently and rent the surplus to other irrigators, thus reducing return flow pollution. It is shown that a water rental market could increase water quality in the Yakima River in southcentral Washington by 31 percent as well as increase farm incomes and crop production.