Abstract. Reservoir is one of the emergency environments that required fast an accurate decision to reduce flood risk during heavy rainfall and contain water during less rainfall. Typically, during heavy rainfall, the water level increase very fast, thus decision of the waterrelease is timely and crucial task. In this paper, intelligentdecisionsupportmodelbased on neuralnetwork (NN) is pro- posed. The proposed model consists of situation assessment, forecasting and decision models. Situation assessment utilized temporal data mining technique to extract relevant data and attribute from the reservoir operation record. The forecasting model utilize NN to perform forecasting of the reservoirwater level, while in the decisionmodel, NN is applied to perform classification of the cur- rent and changes of reservoirwater level. The simulations have shown that the performances of NN for both forecasting and decision models are acceptably good.
Backpropagation networkmodel is a learning technique or training supervised learning the most widely used. This method is one method that is excellent in dealing with complex patterns of recognition. In backpropagation network, each unit that is associated with each input layer units in the hidden layer. Each unit in the hidden layer is connected to every unit in the output layer. This network consists from the many layers (multilayer network). When the network is given input pattern as the training pattern, so the pattern to the hidden layer units to be forwarded to the next layer of output units. Then the output layer units will give you a response as artificial neuralnetwork output. When the output is not as expected, so output will be propagated backward in the hidden layer and then from the hidden layer to the input layer. (Kusumadewi, 2004) 2.3 Fuzzy Multi-Attribute Decision Making
First and foremost, I sincerely thank my parents, to my late father Abdul Mokhtar Ahmad and my mother Siwah Mat Long. Thanks for your support and understanding. Without both of you, I will have never been who I am. To the most important person in my life, my husband, Mursyidul ‘Azim Mazlan, thank you for the love, support, sacrifices and for being with me all the time.
Data was imported into MS Excel and sorted based on the date. A column that represents gate opening/closing was clean to remove noise. Gate opening/closing value is in range of zero to six. Zero indicates gate is closed and values from one to six indicate the number of gates that are open. The change of this value implies the decision point. At this point window slice will be formed begin from that point and preceding to w days according the window size. In this study, the segmentation processes based on sliding window technique begin with window size 2, that represent 2 days of delay. The maximum window size was set to 10. Each segmentation process will return a total of 124 instances. Redundant and conflicting instances are then removed. Table 2 shows the usable number of instances and the window size.
c) ease the exploration of options and changing assumptions for a given problem, d) transference of expertise between one level in the organization to different levels is improved, e) can be used as training tools, f) building a knowledge base helps to establish generalizations, identify gaps and inconsistencies in current knowledge and provide a stimulus for further research. NeuralNetwork is an algorithm that dynamically inherits human neuron information processing capability . This capability enables NeuralNetwork to perform a brain like function such as forecasting, classification, and pattern matching. Genetic Algorithm is an evolutionary algorithm that inspired by natural selection and natural genetics . Genetic Algorithm being used to solve problems related to numerical optimisation. It has been applied in wide range of application including image processing, medicine, robotics, water networks, job scheduling and control.
decisionsupport tools to the pharmacist. In fact, some have gone so far as to suggest a focus on Information Technology based solutions (Avorn 2001). A system of this nature is termed an Expert or Knowledge-Based System, and should help to improve consistency and quality of service, reassure reviewers of their conclusions and if possible speed up the process of performing the review. Furthermore, the system should be easily and naturally maintained and alterable, since new information is being added to the domain on a regular basis. To date it is not believed that such a system exists commercially, with medication review software focusing mostly on data-storage and reporting facilities, providing no heuristic support or reasoning capabilities (Kinrade 2003). Even in research, systems of this nature are fraught with problems, classically being extremely difficult to keep reliably maintained for long periods of time over extensive domains (Compton & Jansen 1989). The closest attempt at a heuristic Knowledge Based System in this vein was developed by Classen et al. for a hospital in the U.S. showing success in detecting sub-optimal drug usage in the hospital environment, but this system is not considered suitable for the broader task of assisting in the general role of medication review (Jha et al. 1998). However, a large, maintainable Knowledge Based System has been contrived in a related domain in the form of LabWizard, a pathology reporting system. In a personal communication from Pacific Knowledge Systems (PKS) this system was shown to demonstrate “over a 29 month period, over 16,000 rules have been added and 6,000,000 cases interpreted with a correct classification rate in the order of 99%” (PKS 2005). It achieved these figures using the Ripple-Down Rules approach to building Knowledge Based Systems, which allows maintenance to become an unobtrusive part of the process of using the system.
As described in previous sections, a case study is performed to test this system on the overall development procedures of the product under study: “study on cloud”. It implemented a concrete architecture of a decisionsupport system for software release management; and assessed the impact of the solution system on management tasks. Quality, progress and health are the major concerns in the present era to evaluate the success or failure of the deployed release (say, better quality, progress and health, means better release productivity). Now providing the prioritization and configuration of these three major concerns in the development and release management of “study on cloud”, leadership team is expected to choose precise solution in problem situation that will help them deploy the best release product. Following this approach, this research claims to provide the greater no. of successful
The paper presents descriptive details of a DSS that has been developed to assist managers of WDN, especially where water supply is intermittent and demand driven and not pres- sure driven. Water distribution networks operate for a limited duration, and because the amount of water each user is able to collect depends on the available pressure at their connection and the duration of the service, the user demand is often not fully met. To improve the service standards, water utilities are struggling hard to develop aids in the form of computer based analysis and management software. Expert systems, as an artificial intelligence (AI) based approach, is one such ap- proach that managers of water utilities are increasingly grav- itating towards. The suggested tool is but a computer pro- gram that seeks to provide solutions to the same problems which were traditionally solved only by a human expert by archiving knowledge in the form of rules and other heuris- tics and then utilizing this pre-programmed set of strategies and the knowledge base in a manner that is akin to human reasoning. The AI field has always fascinated researchers who have applied it in many diverse fields (Cheng et al., 2002; Wang et al., 2004; Muttil and Chau, 2007). In the field of water management, some important applications of expert systems include EXPLORE (Leon et al., 2000), OA- SIS (Goforth and Floris, 1991), CRITQUING Expert System (Shepherd and Ortolano, 1996), IITWSEXP (Khosa et al., 1995), Expert System for treated water distribution (Bunn and Helms, 1999), Network Management System For Water Distribution System (Raghvendran et al., 2007), and Intel- ligent Control System For a Municipal Water Distribution Network (Chan et al., 1999). Most of these initiatives in- cluded components to transfer knowledge from the heuristic domain to the knowledge base of the expert system, while some of the other developments have applied fuzzy logic to process information and suggest “best practice” guidelines for the network manager.
Credit scoring is regarded as a business process to aid the decision of lenders, in this case financial institutions, whether or not lend money to client, assess the risk of their investment and ensure the maximum profit for the loan providers. A simple definition for the credit scoring is given by Anderson (15) which defines it as “(...) the use of statistical models to transform relevant data into numeric measures that guide credit decisions”. This definition considers only the model creation and its influence on credit decision. However, more complete definitions were also found in the literature, which detail the influence of this process on credit decisions. Consequently, Thomas, Elderman and Crook in (16) define the process of credit scoring as “(…) a set of decision models and their underlying techniques that aid lenders in granting consumer credit. These techniques decide who will get credit, how much credit should they get, and what operational strategies will enhance the profitability of the borrowers to the lenders”. This definition clearly states that credit scoring is more than the traditional client assessment view. It also develops considerations about risk, profit and the use of different models and techniques in the decision process.
In quest of emulating the working principles of human brain, ANN is a field of computational science developed over the recent years to deal with the complex systems, which are very difficult or even impossible to model using other analytical and statistical methods. ANN is well suited for prediction purpose as it can approximate successfully any measurable function . The forecasting capabilities of ANN were acknowledged in the past . ANN shows great adaptability, robustness and fault tolerance due to the large number of highly interconnected processing elements . The surface ﬁtting capabilities of ANNs will be essential for our case and this is the reason why ANN was selected to be used for our study .
Reducing fuel consumption of ships against volatile fuel prices and greenhouse gas emissions resulted from international shipping are the challenges that the industry faces today. The potential for fuel savings is possible for new builds, as well as for existing ships through increased energy efficiency measures; technical and operational respectively. The limitations of implementing technical measures increase the potential of operational measures for energy efficient ship operations. Ship owners and operators need to rationalise their energy use and produce energy efficient solutions. Reducing the speed of the ship is the most efficient method in terms of fuel economy and environmental impact. The aim of this paper is twofold: (i) predict ship fuel consumption for various operational conditions through an inexact method, Artiﬁcial NeuralNetwork ANN; (ii) develop a decisionsupport system (DSS) employing ANN based fuel prediction model to be used on-board ships on a real time basis for energy efficient ship operations. The fuel prediction model uses operating data -‘Noon Data’ - which provides information on a ship’s daily fuel consumption. The parameters considered for fuel prediction are ship speed, revolutions per minute (RPM), mean draft, trim, cargo quantity on board, wind and sea effects, in which output data of ANN is fuel consumption. The performance of the ANN is compared with multiple regression analysis (MR), a widely used surface ﬁtting method, and its superiority is conﬁrmed. The developed DSS is exemplified with two scenarios, and it can be concluded that it has a promising potential to provide strategic approach when ship operators have to make their decisions at an operational level considering both the economic and environmental aspects.
Abstract — Arsenic contamination of groundwater in many nations including Bangladesh shows that this is a global problem. Because of the delayed health effects, poor reporting, and low levels of awareness in some communities, the extent of the adverse health problems caused by arsenic in drinking water is at alarming level in Bangladesh. Also, allocating resources such as tube wells efficiently and effectively to mitigate arsenic hazard is a challenging task in Bangladesh. To allocate resources based on different arsenic hazard parameters, we have developed a DecisionSupport System that enables the user to observe the effect of allocation policy both in tabular and spatial format using statistical models. We have also developed an algorithm for optimal allocation of resources. A Smart User Interface is designed for the users so that they will find an interactive, user-friendly, intelligible, logical, clear, and sound environment to work with. Finally, we have analyzed and demonstrated the efficacy of our algorithm graphically.
The baseline scenario, representing business as usual for the catchment was modeled in SWAT. A typical cotton growth cycle included: a) crop sowing at the end of April, b) 190 kg of Nitrogen and 35 kg of Phosphorus fertilization per ha, c) 10 irrigation operations with a 50 mm dose applied from the end of May until the end of August at a 10-days interval, d) harvest in September with crop yields ranging between 3-4 tn ha -1 and e) a soil tillage operation in early November. For a big part of the area under study, irrigation water was provided from the Plastiras artificial lake located outside the catchment (Figure 1). In order to accelerate the development and first implementation of our tool described in this study, water was provided from an external unlimited source of water to the whole agricultural land, ignoring possible groundwater abstractions within the catchment. Complete time- series of measured precipitation and temperature were provided by both the Public Power Corporation (PPC) of Greece and the Hellenic Ministry for the Environment, Physical Planning and Public Works (MEPPPW) for several stations (Figure 1). Annual precipitation ranged from 600 mm in the eastern part of the catchment to more than 1500 mm in the most western edge. Observed monthly river flows at the outlet were also provided for a 27-year period (1970-1996) by MEPPPW. Detailed information on soil properties was not available for the Ali Efenti catchment, thus the three major geological types of flysch, limestones and alluvial were corresponded to an impermeable, a permeable and a semi-permeable soil respectively, based on knowledge on their hydrological behavior. Potential evapotranspiration was calculated with the Penman-Monteith method using historical mean monthly values for meteorological variables. This model setup along with the adjustment of a few soil parameters was found to represent adequately catchment runoff responses. The comparison with the observed monthly flows led to Nash-Sutcliffe coefficients (Nash and Sutcliffe, 1970) above 0.7 for both the calibration and validation periods. Finally, cotton yields were estimated 3.8 tn ha -1 on average, comparable to the actual yields harvested each year. A more detailed description of the Ali Efenti hydrological model parameterization and calibration procedures can be found in the older study of Panagopoulos et al. (2011).
The PNN was first pioneered by Specht (1990). PNN is also a category of feed-forward neuralnetwork which consists of three layers of neurons—namely, input, a hidden layer containing radial basis, and lastly, competitive. The first layer feed inputs data to the hidden layer neurons. Distances among inputs to the training vectors are computed by radial basis function. The computed results indicate the closeness of input to a training input. The last layer of the PNN summed the contributions produced by every class of input to produce the PNN probabilities output. At the last layer, the competitive transfer function classifies the inputs due to its optimum probability of classification accuracy (Mantzaris et al., 2011). In contrast to other types of NN, PNNs are only applicable in solving classification problems, and the majority of their training techniques are easy to use. Assuming PNN accepts a vector (x) as input, x ( x 1 ,..., x n ) is applied to neurons in input layer x i ( 1 i n ) and is
This paper presents a multi-criteria decision making (MCDM) model for evaluating an Information and Communication Technology (ICT) network system in health care .The competing goals existing in Health Institutions need a special treatment, thus the MCDM approach is essential for identifying ICT network quality of service (QoS) requirements and implications. A pilot study based on user perception is explored involving three categories of hospitals in Chile. Data is collected considering various health sector representatives. The main contribution is the proposed decision methodology to develop criteria for evaluating QoS issues of an ICT network system within a healthcare environment using the Analytical Hierarchy Process (AHP). The results provides a framework to make decisions concerning an information technology networked system, characterizing end users and their needs and enabling tradeoffs in agreement with the institution objectives.
III. S USTAINABILITY OF B USINESS D EVELOPMENT With environmental and social change combining with the increasingly competiveness of the global economy, sustainabil- ity of businesses [17,18] is a key issue identified for the Dar- ling Downs region . The Darling Downs region in Queens- land, Australia, is a prominent food production and mineral resource producing area . Individual business and govern- ment organisations need to consider decisions and policies made in light of the interdependence between themselves and the wider community. This complexity calls for new methods or models in the process of decision making in the future to be sustainable. The concept of regional sustainability is illustrated in Fig.3.
Based on the experience from previous Integrated Mobile Information System (IMIS), a new web based system, IDSS has been developed. The scope of this web based application covers type 2 diabetes patients as well as clinicians. Diabetes type 2 patients will use IDSS for their self-management regarding their diet and exercise and clinicians will use IDSS to give advises on diet and exercise. In the patient perspective IDSS behave as a tool for education and communication. The important part in the education of diabetes patient is learning how to attain achieve glycaemic. Another aspect of the IDSS is the learning environment. Patients can experiment with their own data adjusting diet and/or exercises routines and learn how to deal with different health conditions. The IDSS also educate patients that how to achieve glycaemic control. In this way different dietary and exercise alternatives can be taken into consideration without taking the risk of actually experiencing hyperglyceamic or hypoglyceamic situations. While experiencing the IDSS patients can learn about the effect calories, exercises routine and carbohydrate intakes. For example patient can learn about the effect of increased exercise in summer vocations. Furthermore as comparing to other traditional diabetes educational materials the major benefit of IDSS is, it utilizes the patient’s own data for diabetes management. In addition the system gives empowerment to patient and enables new ways of self-treatment.
There is an increasing interest in combination of RBR and CBR, and integration of other reasoning methods such as Bayesian Belief Network (BBN) in developing IDSS in recent years. BBNs are probabilistic inference engines that can be used to reason under uncertainty. There is plenty of ongoing research on integrating BBNs into a wide range of decision making fields especially to solve complex semi structured and unstructured decision problems. For example, Lauria and Duchessi (2006) discussed how to create a BBN from real world data on information technology implementation and how it was incorporated into an IDSS to support “what if” analysis. Anderson and Vastag (2004) explored how to use BBN for causal modelling in operations research. Kristensen and Rasmussen (2002) reported how to build a BBN for decisionsupport in agriculture. Approaching from a slightly different angle, Lin et al (2007) discussed employing fuzzy set theory for decision making in selection data warehouse systems for enterprises, particularly on decision alternatives with decision criteria. The idea is tested through a prototype IDSS with a case study in agriculture. The system demonstrates that it is easier for decision makers to collect data, calculate data, and to interpret results (i.e. automatic ranking order of the alternatives) through utilisation of triangular fuzzy numbers, therefore the system improves the decisions by considering the vagueness, ambiguity and uncertainty prevalent in real word systems. Within above mentioned work, specific implementation of the integration solutions within specific IDSS may not be all the same, but there are three common steps in terms of model integration: search for candidate models, benchmark candidate models and apply the selected model. It is also clear that the hybrid approach greatly facilitates the model integration under dynamic and uncertain decision situations. As the reasoning methods like BBN and fuzzy theory represent models with causal relationships among a set of variables of interest, a set of conditional independence assumptions, and their related joint probabilities, and the variables assumptions probabilities are defined for specific applications, the integration involved in the BBN and fuzzy theory supported IDSS are often tight, in project integration in nature.
Abstract. The paper describes the design of a genetic classifier-based intrusion detection system, which can provide active detection and automated responses during intrusions. It is designed to be a sense and response system that can monitor various activities on the network (i.e. looks for changes such as malfunctions, faults, abnormalities, misuse, deviations, intrusions, etc.). In particular, it simultaneously monitors networked computer’s activities at different levels (such as user level, system level, process level and packet level) and use a genetic classifier system in order to determine a specific action in case of any security violation. The objective is to find correlation among the deviated values (from normal) of monitored parameters to determine the type of intrusion and to generate an action accordingly. We performed some experiments to evolve set of decision rules based on the significance of monitored parameters in Unix environment, and tested for validation.
In the future economy, a knowledge<based economy, decisionsupport systems (DDS) are very rigorous and precise, if the hypothesis is well grounded. An important direction of research is simulation of specialist thinking based on a Knowledge Based Systems (KBS). The evolution of DDS and KBS depends on the evolution of knowledge representation. Even though the researches in economic knowledge representation are in progress, the cases in which the theory is put into practice are very rare, and of limited complexity. An evolved KBS must incorporate knowledge pieces capable of explaining the economic phenomenon in all it’s complexity. In the near future not only the problem of rational, conscious knowledge will be a problem, but the one of unconscious knowledge based on intuition and imagination. All of these will be a support for the development of economic axiomatic systems.