Evaluation of reservoir sedimentation using data driven techniques

The sedimentation is a pervasive complex hydrological process subjected to each and every reservoir in world at different extent. Hydrographic surveys are considered as most accurate method to determine the total volume occupied by sediment and its distribution pattern in a reservoir. But, these surveys are very cumbersome, time consuming and expensive. This complex sedimentation process can also be simulated through the well calibrated numerical models. However, these models generally are data extensive and require large computational time. Generally, the availability of such data is very scarce. Due to large constraints of these methods and models, in the present study, data driven approaches such as artificial neural networks (ANN), model trees (MT) and genetic programming (GP) have been investigated for the estimation of volume of sediment deposition incorporating the parameters influenced it along with conventional multiple linear regression data driven model. The aforementioned data driven models for the estimation of reservoir sediment deposition were initially developed and applied on Gobindsagar Reservoir. In order to generalise the developed methodology, the developed data driven models were also validated for unseen data of Pong Reservoir. The study depicted that the highly nonlinear models ANN and GP captured the trend of sediment deposition better than piecewise linear MT model, even for smaller length datasets.     

A novel artificial bee colony algorithm with Powell's method

Artificial bee colony (ABC) algorithm is a relatively new optimization technique which has been shown to be competitive to other population-based algorithms. However, there is still an insufficiency in ABC regarding its solution search equation, which is good at exploration but poor at exploitation. To address this concerning issue, we first propose a modified search equation which is applied to generate a candidate solution in the onlookers phase to improve the search ability of ABC. Further, we use the Powell's method as a local search tool to enhance the exploitation of the algorithm. The new algorithm is tested on 22 unconstrained benchmark functions and 13 constrained benchmark functions, and are compared with some other ABCs and several state-of-the-art algorithms. The comparisons show that the proposed algorithm offers the highest solution quality, fastest global convergence, and strongest robustness among all the contenders on almost all test functions.     

图灵机的哲学意义

人类在创造人工自然的过程在中,如果不遵循自然规律而一意孤行,其结果必然对天然自然造成级大破坏,人类最终将自食苦果,计算机技术的发展也不例外。     

Augmented Panoramic Video

Many video sequences consist of a locally dynamic background containing moving foreground subjects. In this paper we propose a novel way of re‐displaying these sequences, by giving the user control over a virtual camera frame. Based on video mosaicing, we first compute a static high quality background panorama. After segmenting and removing the foreground subjects from the original video, the remaining elements are merged into a dynamic background panorama, which seamlessly extends the original video footage. We then re‐display this augmented video by warping and cropping the panorama. The virtual camera can have an enlarged field‐of‐view and a controlled camera motion. Our technique is able to process videos with complex camera motions, reconstructing high quality panoramas without parallax artefacts, visible seams or blurring, while retaining repetitive dynamic elements.     

The development of a formative scenario-based computer assisted assessment tool in psychology for teachers: The PePCAA project

Formative computer assisted assessment has become increasingly attractive in Higher Education where providing useful feedback to large numbers of students can be difficult. However, the nature of such assessments has often been limited to objective questions such as multiple-choice. This paper reports on the development and initial trialling of a more innovative, formative use of computer assisted assessment in a Higher Education context. The European funded PePCAA (Pedagogical Psychology Computer Assisted Assessment) project developed a series of scenario-based computer-delivered formative assessments of pedagogical psychology for teachers and trainee teachers, using a range of software features, including the addition of confidence measurement. The project had a two-fold aim: to provide a tool to improve understanding of pedagogical psychology and to explore the potential of more innovative techniques of computer assisted assessment to motivate students and to assess deeper learning. The combination of computer-delivered formative assessment with innovative question styles and confidence ratings is believed to be unique for pedagogical psychology. Scenarios were based on realistic classroom situations and focused on problem solving or on utilising best practice. The PePCAA Learning Assessment Circle (PLAC) provided a framework for indexing the kinds of processes required of users. In the UK, small scale trialling involved a total of 23 teacher trainees such that each assessment was attempted by about seven participants. Participants completed evaluation questionnaires after each assessment. Responses from learners indicated that the UK scenarios were generally very well received and had at least partly achieved the aim of stimulating deeper learning. Transfer of assessments between countries proved more difficult than expected. The next stage of development should be to conduct a larger pilot, thus allowing full investigation of the reliability and validity of the assessments. There is also scope for further development of the PePCAA approach and for its application in other subjects.     

Forming processes control by means of artificial intelligence techniques

Forming processes are manufacturing processes that use force and pressure in order to modify the shape of a material part until obtaining the final product. The wide range of non-linear factors that drive this sort of processes make them very complex and extremely difficult to be controlled. Traditional control techniques, like PID controllers, have not offered a reliable solution when global control has been pursued and the figure of the operator still remains present in most of the forming facilities. On the other hand, although operators have demonstrated to be a very successful strategy when controlling this type of processes, the actual market evolution towards the fabrication of more complex parts, made of lower formability materials at higher production rates, is decreasing their capacity of reaction when solving the daily problems. Thus, the development of new global control systems based not on traditional control techniques and mathematical models but on the control strategy that has been used successfully for many years, the control through the experience and knowledge is now even more necessary. In the present work, an intelligent control system based on one of the main techniques within the artificial intelligence, expert systems, has been developed. The main purpose of this intelligent control system is to emulate the decisions that expert operators take but in a quicker and more reliable way. The developed intelligent control system has been installed in a blanking facility and very good results have been achieved.     

Inductive transfer with context-sensitive neural networks

Context-sensitive Multiple Task Learning, or csMTL, is presented as a method of inductive transfer which uses a single output neural network and additional contextual inputs for learning multiple tasks. Motivated by problems with the application of MTL networks to machine lifelong learning systems, csMTL encoding of multiple task examples was developed and found to improve predictive performance. As evidence, the csMTL method is tested on seven task domains and shown to produce hypotheses for primary tasks that are often better than standard MTL hypotheses when learning in the presence of related and unrelated tasks. We argue that the reason for this performance improvement is a reduction in the number of effective free parameters in the csMTL network brought about by the shared output node and weight update constraints due to the context inputs. An examination of IDT and SVM models developed from csMTL encoded data provides initial evidence that this improvement is not shared across all machine learning models.     

Enacting agent-based services for automated procurement

Negotiation events in industrial procurement involving multiple, highly customisable goods pose serious challenges to buying agents when trying to determine the best set of providing agents’ offers. Typically, a buying agent's decision involves a large variety of constraints that may involve attributes of a very same item as well as attributes of different, multiple items. In this paper we present iBundler, an agent-aware service offered to buying agents to help them determine the optimal bundle of received offers based on their business rules. In this way, buying agents are relieved with the burden of solving too hard a problem and concentrate on strategic issues. iBundler is intended as a negotiation service for buying agents and as a winner determination service for reverse combinatorial auctions with side constraints. Furthermore, we assess the computational cost added by employing agent technology in the development of iBundler to characterise the type of negotiation scenarios that it can acceptably handle.     

Optimized scenario for rainfall forecasting using genetic algorithm coupled with artificial neural network

Rainfall forecasting plays many important role in water resources studies such as river training works and design of flood warning systems. Recent advancement in artificial intelligence and in particular techniques aimed at converting input to output for highly nonlinear, non-convex and dimensionalized processes such as rainfall field, provide an alternative approach for developing rainfall forecasting model. Artificial neural networks (ANNs), which perform a nonlinear mapping between inputs and outputs, are such a technique. Current literatures on artificial neural networks show that the selection of network architecture and its efficient training procedure are major obstacles for their daily usage. In this paper, feed-forward type networks will be developed to simulate the rainfall field and a so-called back propagation (BP) algorithm coupled with genetic algorithm (GA) will be used to train and optimize the networks. The technique will be implemented to forecast rainfall for a number of times using rainfall hyetograph of recording rain gauges in the Upper Parramatta catchment in the western suburbs of Sydney, Australia. Results of the study showed the structuring of ANN network with the input parameter selection, when coupled with GA, performed better compared to similar work of using ANN alone.     

An integrated multi-objective immune algorithm for optimizing the wire bonding process of integrated circuits

Optimization of the wire bonding process of an integrated circuit (IC) is a multi-objective optimization problem (MOOP). In this research, an integrated multi-objective immune algorithm (MOIA) that combines an artificial immune algorithm (IA) with an artificial neural network (ANN) and a generalized Pareto-based scale-independent fitness function (GPSIFF) is developed to find the optimal process parameters for the first bond of an IC wire bonding. The back-propagation ANN is used to establish the nonlinear multivariate relationships between the wire boning parameters and the multi-responses, and is applied to generate the multiple response values for each antibody generated by the IA. The GPSIFF is then used to evaluate the affinity for each antibody and to find the non-dominated solutions. The “Error Ratio” is then applied to measure the convergence of the integrated approach. The “Spread Metric” is used to measure the diversity of the proposed approach. Implementation results show that the integrated MOIA approach does generate the Pareto-optimal solutions for the decision maker, and the Pareto-optimal solutions have good convergence and diversity performance.