The multi-purpose reservoirwaterreleasedecision requires an expert to make a decision by assembling complex decision information that occurred in real time. The decision needs to consider adequate reservoirwater balance in order to maintain reservoir multi-purpose function and provide enough space for incoming heavy rainfall and inflow. Crucially, the waterrelease should not exceed the downstream maximum river level so that it will not cause flood. The rainfall and water level are fuzzy information, thus the decisionmodel needs the ability to handle the fuzzy information. Moreover, the rainfalls that are recorded at different location take different time to reach into the reservoir. This situation shows that there is spatialtemporal relationship hidden in between each gauging station and the reservoir. Thus, this study proposed dynamicreservoirwaterreleasedecisionmodel that utilize both spatial and temporal information in the input pattern. Based on the patterns, the model will suggest when the reservoirwater should be released. The model adopts Adaptive Neuro-Fuzzy Inference System (ANFIS) in order to deal with the fuzzy information. The data used in this study was obtained from the Perlis Department of Irrigation and Drainage. The modified Sliding Window algorithm was used to construct the rainfall temporalpattern, while the spatial information was established by simulating the mapped rainfall and reservoirwater level pattern. The model performance was measured based on the Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). Findings from this study shows that ANFIS produces the lowest RMSE and MAE when compare to Autoregressive Integrated Moving Average (ARIMA) and Backpropagation Neural Network (BPNN) model. The model can be used by the reservoir operator to assist their decision making and support the new reservoir operator in the absence of an experience reservoir operator.
Reservoir is one of major structural approaches for flood mitigation. During floods, early reservoirwaterrelease is one of the actions taken by the reservoir operator to accommodate incoming heavy rainfall. Late waterrelease might give negative effect to the reservoir structure and cause flood at downstream area. However, current rainfall may not directly influence the change of reservoirwater level. The delay may occur as the streamflow that carries the water might take some time to reach the reservoir. This study is aimed to develop a forecasting model for the change in stage of reservoirwater level. The model considers the changes of reservoirwater level and its stage as the input and the future change in stage of reservoirwater level as the output. In this study, the Timah Tasoh reservoir operational data was obtained from the Perlis Department of Irrigation and Drainage (DID). The reservoirwater level was categorised into stages based on DID manual. A modified sliding window algorithm has been deployed to segment the data into temporal patterns. Based on the patterns, three models were developed: the reservoirwater level model, the change of reservoirwater level and stage of reservoirwater level model, and the combination of the change of reservoirwater level and stage of reservoirwater level model. All models were simulated using neural network and their performances were compared using on mean square error (MSE) and percentage of correctness. The result shows that the change of reservoirwater level and stage of reservoirwatermodel produces the lowest MSE and the highest percentage of correctness when compared to the other two models. The findings also show that a delay of two previous days has affected the change in stage of reservoirwater level. The model can be applied to support early reservoirwaterreleasedecision making. Thus, reduce the impact of flood at the downstream area.
Abstract—Emergency situation required fast and accurate decision as every decision is very critical to save human lives. Naturally, during this situation humans made decisionbased on their past experiences by which their nerves and brain system will perceive the situation and mapped with their experiences to produce action. This naturalistic decision making approach has been one of the attention in emergency management research. In this paper a conceptual model of Intelligent Decision Support System for reservoir operation during emergency situation is proposed. This model simulates human decisionbased on three models: situation assessment, expectancy forecasting, and decision modeling. Situation assessment is to extract the temporal data from the hydrological and operational databases. This data will be used in the forecasting module, to forecast the future event. The decision module will utilized the temporal and the forecasted data to produce the final decision. Artificial intelligence techniques are utilized in every model. The model is expected to assist reservoir operator in making decision during emergency situation; typically during heavy rainfall when early and fast decision is required to release the reservoirwater in order to leave enough space for incoming water and to release the water in the save carrying capacity of the downstream channel. Thus avoiding flood in downstream areas.
Reservoirwaterreleasedecision is one of the challenging tasks for reservoir operators in order to determine the amount of water to be stored and to be released from a reservoir (Norwawi et al., 2005; Wurbs, 1993). The reservoir capacity needs to be maintained in order to prevent downstream floods and to reduce water shortage problems in the future. In both flood and drought situations, the decisions regarding waterrelease are made in accordance with the available water, inflows, demands, time, previous releases, etc. (Jain & Singh, 2003). However, different reservoirs have different objectives and purposes, thus different operation rules are needed (Wan Ishak et al., 2012). Typically, reservoirwaterreleasedecision is based on upstream inflow that is observed through the magnitude of the upstream rainfall and the river water level. The total volume of rainfall may come from several gauging stations and their distances to the reservoir are varied (Mokhtar et al., 2016). Thus, rainfall observed at those gauging stations may take different time to reach the reservoir. This situation shows that there is a spatialtemporal relationship hidden between each gauging station and the reservoir. Currently, there are limited studies that focus on this situation, as most of the previous studies focused on temporal relationship and the total rainfall volume (Mohan & Revesz, 2012; Wan Ishak et al., 2012; Mokhtar et al., 2014). Thus, the spatialtemporal relationship needs to be considered for modelling reservoirwaterreleasedecision.
Abstract: In this paper, a three-dimensional spatial-temporal decomposition modelling method is proposed to build the alkali-surfactant-polymer (ASP) flooding model, in which a new dynamic recurrent wavelet neural network (DRWNN) is presented to identify the temporal coefficients. At first, the detailed mathematical model of ASP flooding is described which is a complex distributed parameter system. Then a three-dimensional spatial-temporal modelling method is inferred based on Karhunen-Loeve (K-L) decomposition to decompose the water saturation of reservoir into a series of spatial basis functions and corresponding temporal coefficients. Furthermore, the recurrent wavelet neural network is used to acquire the identification model, in which the injection concentrations of ASP flooding and temporal coefficients are taken as the input and output information. In order to improve the capability of dynamic modelling, DRWNN is proposed through adding feedback layers and setting the different weights with time to achieve dynamic memory of the past information. Considering the gradient descent method for the neural networks training easily leads to local minimum and slow convergence speed, the spectral conjugate gradient method is introduced to optimize the weights of DRWNN. At last, DRWNN is used to build the relation between the moisture content of production wells and the water saturation of the corresponding grids. Thus, the final approximate model of ASP flooding is finished. The accuracy is proved by model with four injection wells and nine production wells through data from the mechanism model.
In this study, nine neural network models were developed. Each neural network model is trained with one data set. Inputs of all data sets are normalized using min- max method and the output was represented based on Binary-Coded-Decimal (BCD) scheme. Each model is trained with different combination of hidden unit, learning rate and momentum. The training is control by three conditions (1) maximum epoch (2) minimum error, and (3) early stopping condition. Early stopping is executed when the validation error continue to arises for several epochs . Fig. 3 shows the procedure for the neural network training. The aim of this procedure is to get the combination that gives the best result. Prior to the training, the each data set is randomly divided into three different sets: training (80%), validation (10%) and testing (10%) sets (Table 3).
The combine implementation of structural and non-structural approach is vital to avoid casualty and false sense of security at storage . Structural approach such as dam is a solid structure that holds water for a certain period or at maximum reservoirwater level. In practice, the waterrelease or the gate opening decision depends on the operating rules . These rules are static and do not consider the dynamic nature of the hydrology systems. Therefore, non-structural approach such as forecasting is vital to support the waterrelease or the gate opening decision. The dynamic of the fore- casting system will be able to cope with the event frequency and triggered alert to the authority when the situation is at the severe level. Flood forecasting is significant to cope with the great floods .
building these models were described in the papers in which the models were introduced, along with a discussion on potential applications. However, the predicted applications are seldom re-examined. As SIMDEUM, a stochastic end-use model for drinking water demand, has often been applied in research and practice since it was developed, we are re- examining its applications in this paper. SIMDEUM’s original purpose was to calculate maximum demands in order to be able to design self-cleaning networks. Yet, the model has been useful in many more applications. This paper gives an 15
Quirijns et al., 2013b; Agudelo-Vera et al., 2013a; Pieterse- Quirijns and Beverloo, 2013) for which insight into the max- imum daily drinking water demand and the maximum daily hot water demand, together with a requirement for the mini- mum and maximum flow velocities, leads to a certain pipe diameter; the maximum demand for hot water in 10 min, 1 h or 1 day is used to determine the most appropriate hot water device. The design rules following from SIMDEUM were validated (see Sect. 2.2, Fig. 6) and are now in the of- ficial Dutch guidelines for the design of DWIs in apartment buildings and non-domestic buildings (ISSO-kontaktgroep, 2015). A research project was also conducted for the design of hot water devices in residential DWIs (Pieterse-Quirijns et al., 2015). This study showed that hot water devices can be designed much smaller, thus reducing the waste of en- ergy, when realistic water demands are taken into account. The current Dutch building regulation is based on floor space only; SIMDEUM can complement this as it can also take into account the presence and behaviour of the residents.
The development of personalised decision support systems has the potential to be the tool for better understanding health related problems like chronic disease includ- ing stroke, cardiovascular disease, cancer and countless unsolved medical problems. For instance, health related problems like chronic diseases are the major cause of death in almost all countries and it is projected that 41 million people will die of a chronic disease by 2015 unless urgent action is taken [Organization 2005]. Vari- ous initiatives have been taken to control the progression of symptoms in chronic disease patients such as clinical prevention using combination of drug therapy and calculation of a person’s risk by referring to an existing risk chart which takes into account several risk factors. Additional initiatives involve the use of statistical meth- ods to generate a survival model and to investigate several risk factors associated with chronic disease, such as the Cox Proportional Hazards Model [Lumley 2002], [Wolf 1991], [Yusuf 1998]. There are also several machine learning applications that used global models for prediction of a person’s risk or the outcome of a certain diseases [Khosla 2010], [Das 2003], [Anderson 2006], [Levey 1999]. According to [Shabo 2007] there is evidence that prediction and treatment based on global mod- els are only effective for some patients (about 70% average) leaving the remaining 30% of patients without proper treatment which could worsen their condition and possibly lead to their death. A global model is derived from all available data for the target and then applied to any new patient anywhere at any time. While it may give 70% to 80% average accuracy over the whole population, it still may not be suitable for many individuals [Kasabov 2010b]. Hence, using global models for pre- diction of a person‘s risk is inadequate, based on the assumption that every person or individual has their own unique characteristics.
healthcare monitoring [7, 8], surveillance [9, 10], and target group pattern generation [11, 12]. Multi-sensing systems outperform traditional tracking techniques in managing the flow of signal and coordinating sensor ac- tions . A tracking system equipped with sensors, which functions in an autonomous or semi-autonomous mode, has to be capable of defining the target from the sensors as well as making decisions based on this infor- mation . The information of different modalities can be applied to obtain an exact understanding of the target with signal integration. In a multi-sensing system, mul- tiple sensors inevitably introduce problems of data re- dundancy or data conflicting due to the data stored in multiple databases with inconsistent attributes [15–17]. The use of available data, the connectivity and diversity between data sources, and abilities of data analytic methods make the tracking precision still a challenge. Without the strategies of precise positioning, the address of sensor output can hardly be reliable. That is, the ac- curacy improvement of sensing system is identified as a specific issue in tracking optimization. The importance of this issue is that it is an underlying concern of the sensing system.
and ˆ θ = 0.123mm/min. The sample and fitted statistics are shown in Table 1 and Figure 2, where it can be seen a reasonable fit is obtained, but with a slight over- estimation of skewness and autocorrelation (Table 1). There was also some slight under-estimation of the sample cross-correlations for those sites having the greater spatial separation (Figure 2). However, given that these discrepancies were only slight and that a very parsimonious model parametrization had been used, no further improvement in fit was sought.
VII. D EFINITION OF R ELEASE P LANNING P ATTERNS Similar to other software developing patterns, release planning pattern should be well defined to be used in release planning. Patterns employed in release planning are divided into two groups based on their effect: patterns effective on all release planning steps and patterns effective on a certain step. The former group is “release planning patterns” and the latter is “release planning steps’ patterns”. Moreover, as mentioned in release planning customization, some parameters are common among some multiple steps and some others only influence a certain one. Putting all parameters together makes it possible to obtain the set of effective parameters on all release planning steps which are called “effective parameters on release planning”. According to what mentioned above, release planning pattern is displaying a best experience in release planning which:
Abstract—The mixed pixel problem is common in remote sensing. A soft classification can generate land cover class fraction images that illustrate the areal proportions of the various land cover classes within pixels. The spatial distribution of land cover classes within each mixed pixel is, however, not represented. Super-resolution land cover mapping (SRM) is a technique to predict the spatial distribution of land cover classes within the mixed pixel using fraction images as input. Spatial-temporal SRM (STSRM) extends the basic SRM to include a temporal dimension by using a finer-spatial resolution land cover map that pre- or post-dates the image acquisition time as ancillary data. Traditional STSRM methods often use one land cover map as the constraint, but neglect the majority of available land cover maps acquired at different dates and of the same scene in reconstructing a full state trajectory of land cover changes when applying STSRM to time series data. In addition, the STSRM methods define the temporal dependence globally, and neglect the spatial variation of land cover temporal dependence intensity within images. A novel local STSRM (LSTSRM) is proposed in this paper. LSTSRM incorporates more than one available land cover map to constrain the solution, and develops a local temporal dependence model, in which the temporal dependence intensity may vary spatially. The results show that LSTSRM can eliminate speckle-like artifacts and reconstruct the spatial patterns of land cover patches in the resulting maps, and increase the overall accuracy compared with other STSRM methods.
The study area is located in Hebei Province (38.850˚N, 116.000˚E). This shallow lake is disk-shaped with sur- face area 366 km 2 and water depth 1 - 2 m. The lake lies in the middle reaches of the Daqinghe River basin and ultimately discharges into the Bohai Gulf, Yellow Sea. Much of the upstream catchment totaling 31,500 km 2 lies within Baoding municipality. While nine rivers and/or channels flow into Lake Baiyangdian, only two have re- gular but decreasing flows, namely, Fuhe and Juma River . Water systems of Baiyangdian area was shown in Figure 1.
An area in the decidua was observed to be positively stained for pimonidazole protein adducts (Fig. 1A, inset a). Hemosiderin, most commonly found in macrophages, was observed in the basal layer (Fig. 1A, asterisk). This observation is consistent with findings from Kaitu’u- Lino et al. who observed macrophages residing in the basal layer at the time of PWD . Within 4 h of PWD (Fig. 1B) the decidualised cell mass was intensely stained for hypoxia, whilst the basal layer remained unstained. The luminal epithelium was negative for hypoxia (indi- cated by the arrows in Fig. 1B, inset b), whilst the stro- mal cells adjacent to the luminal epithelium were positively stained. The intensity of staining of pimoni- dazole adducts was stronger in the shedding decidual cell mass in tissues collected 8 h after the withdrawal of progesterone (Fig. 1C). The luminal epithelium remained unstained. The basal layer also remained normoxic, with the existence of a defined gradient in the intensity of staining in the decidualised cells and the underly- ing basal layer (Fig. 1C, inset c) consistent with spatial
In this study, the Weather Research and Forecasting (WRF-ARW) model was used to investigate the influence of topography on precipitation in Rwanda, as in . The Weather Research and Forecasting (WRF) model is considered as the latest and widely used mesoscale models which is used by both operational and research communities . The Advanced Research WRF (ARW) is the main component of the WRF modeling system in which there are several initialization programs for idealized and real-data simulations. There is a fully com- pressible, Eulerian and non-hydrostatic equations with a run-time hydrostatic option which are conservative for scalar variables in the ARW dynamical core . The WRF model has the ability to simulate extreme weather events or events of short time . It provides several parameterization options named as microphysics, cumu- lus parameterization, surface layer, land surface model and planetary boundary layer . The WRF model uses the terrain-following, hydrostatic-pressure vertical coordinate with constant pressure surface as top of the model boundary. The third order Runge-Kutta scheme while the spatial discretization employs 2nd- to 6th-order advec- tion options. The model has several initial, lateral and top boundaries  and it can use the interactive one-way, two-way, and moving nesting options .
ABSTRACT: In this paper, we propose orthogonal polynomial based image retrieval using multiresolution wavelet features. Orthogonal polynomial high-order coefficient and low-order coefficient are enables the low-frequency and high-frequency subbands in multiresolution structure. For effective content based image retrieval we used combined these two features. Here polynomial model used to posses the localization frequency information in spatial-temporal domain. Wavelet Packet (WP) transforms is used to construct various subbands in order to extract the feature such as color and texture. For color features we used color autocorrelogram and for texture features we consider the second order statistics in each subbands. Based on these features, feature vector is formed and the experimental result is implemented in two databases. City-block distance method is incorporated to find the distance between query image and target image. The experimental results reveal that the proposed method performs better retrieval accuracy when compared to those existing methods.
Most of the upper stream areas are facing descending precipitation and higher temperature. Over the flow period, the southern part has a larger change rate of temperature than the north, the largest rate is about 0.9 ° C/10-years. And the upper stream areas have significant descending precipitation, the largest rate of which is about –80 mm/10-years. Research by Gu (2011) based on in-field meteorological data reached a similar conclusion by the linear regression method. With higher temperature and less precipitation, the Hailar River basin and upper Argun River area would have less water resources.
The internet has changed people's lives, and tourists have used social media to share information such as photos, microblogs, and travel notes, which has produced a huge amount of tourism data and various forms of structure. Therefore, the development of a reasonable Internet tourism data mining method and the establishment of a spatial-temporal behavior model of the passenger flow in the scenic spot will help to improve the tourism efficiency and safety management of the scenic spot. Andrea and Others (2009) used digital footprint information such as pictures, call logs and mobile text messages to study the 2006-2008-year tourism situation in New York . Van den Berg and Others (2013) studied the impact of social networks on tourism activity patterns . Li Chunming and Others (2013) analyzed the spatial-temporal behavior of tourists in Gulangyu scenic spot based on network photos . Chen Rong and Others (2014) proposed the ssvr-pso passenger flow prediction modelbased on support vector regression and particle swarm optimization _______________________________________________________________