One View of the Future

Abstract

Floods, droughts, water scarcity, and water contamination are some among many water problems that are present today and will be even more noticeable in the future. In the past, many different tools have been used for simulation and optimization of complex water resources systems in order to provide an improved basis for decision making. The continuing evolution of information technology (hardware and software) creates a good environment for the transition to new tools. Application of the systems approach to water resources planning, management, and operations has been established as one of the most important advances made in the field of water resources engineering. Based on the lessons learned, this contribution provides my personal view on the tools to be used in the future. Two paradigm shifts are discussed. The first one is focusing on the complexity of the water resources domain and the complexity of the modelling tools in an environment characterised by continuous rapid technological development. The second one deals with water-related data availability and natural variability of domain variables in time and space affecting the uncertainty of water resources decision making.     

Comparison of Representative Heuristic Algorithms for Multi-Objective Reservoir Optimal Operation

Heuristic algorithms (HAs) are widely used in multi-objective reservoir optimal operation (MOROO) due to the rapidity of the calculation and simplicity of their design. The literature usually focuses on one or two categories of HAs and simply reviews the state of the art. To provide an overall understanding and a specific comparison of HAs in MOROO, differential evolution (DE), particle swarm optimisation (PSO), and artificial physics optimisation (APO), which serve as typical examples of the three categories of HAs, are compared in terms of the development and applications using a designed experiment. Besides, the general model with constraints and fitness function, and the solution process using a hybrid feasible domain restoration method and penalty function method are also presented. Taking a designed experiment with multiple scenarios, the mean average of the optimal objective function values, the standard deviation of optimal objective function values, the mean average of the computational time, and population diversity are used for comparisons. Results of the comparisons show that (a) the problem of optimal multipurpose reservoir long-term operation is a mathematic programming problem with narrow feasible region and monotonic objective function; (b) it is easy to obtain the same optimal objective function value, but different optimal solutions using HAs; and (c) comparisons do not result in a clear winner, but DE can be more appropriate for MOROO.

     

The Use of Object-Oriented Modeling for Water Resources Planning in Egypt

Water problems are omnipresent and are already becoming a limiting factor in the development of many countries. Currently the balance between the available and required water in Egypt is fragile. Any movement away from the balancing point means either less ambitious economic development or depletion of the resources and degradation of the environment. The continuing revolution in computer hardware and software is expected to provide more insight into the water problems and to alleviate some of the future water crises. In this paper we have investigated potential benefits which can be accrued from the application of object-oriented modeling in water resources.     

A modified dual current mode control method with an adaptive current bandwidth

In this paper, a new adaptive dual current mode control method (ADCMC) is presented, being a result of the modification of existing dual current mode control (DCMC) by introducing an adaptive current bandwidth. The ADCMC offers several important advantages over DCMC, such as no peak‐to‐average error in the inductor current, better transient response of current loop, and improved line regulation. A detailed analysis of the proposed ADCMC is performed for three types of DC–DC power electronics converters: buck, boost, and non‐inverting buck–boost converter. The performances of the ADCMC are tested with simulations and experiments. The obtained results confirm the analysis and validity of the proposed ADCMC method. Copyright © 2015 John Wiley & Sons, Ltd.     

Decentralized fault detection and diagnosis via sparse PCA based decomposition and Maximum Entropy decision fusion

This paper proposes an approach for decentralized fault detection and diagnosis in process monitoring sensor networks. The sensor network is decomposed into multiple, potentially overlapping, blocks using the Sparse Principal Component Analysis algorithm. Local predictions are generated at each block using Support Vector Machine classifiers. The local predictions are then fused via a Maximum Entropy algorithm. Empirical studies on the benchmark Tennessee Eastman Process data demonstrated that the proposed decentralized approach achieves accuracy comparable to that of the fully centralized approach, while offering benefits in terms of fault tolerance, reusability, and scalability.     

Parsing human skeletons in an operating room

Multiple human pose estimation is an important yet challenging problem. In an operating room (OR) environment, the 3D body poses of surgeons and medical staff can provide important clues for surgical workflow analysis. For that purpose, we propose an algorithm for localizing and recovering body poses of multiple human in an OR environment under a multi-camera setup. Our model builds on 3D Pictorial Structures and 2D body part localization across all camera views, using convolutional neural networks (ConvNets). To evaluate our algorithm, we introduce a dataset captured in a real OR environment. Our dataset is unique, challenging and publicly available with annotated ground truths. Our proposed algorithm yields to promising pose estimation results on this dataset.     

Supervised clustering of label ranking data using label preference information

This paper studies supervised clustering in the context of label ranking data. The goal is to partition the feature space into K clusters, such that they are compact in both the feature and label ranking space. This type of clustering has many potential applications. For example, in target marketing we might want to come up with K different offers or marketing strategies for our target audience. Thus, we aim at clustering the customers’ feature space into K clusters by leveraging the revealed or stated, potentially incomplete customer preferences over products, such that the preferences of customers within one cluster are more similar to each other than to those of customers in other clusters. We establish several baseline algorithms and propose two principled algorithms for supervised clustering. In the first baseline, the clusters are created in an unsupervised manner, followed by assigning a representative label ranking to each cluster. In the second baseline, the label ranking space is clustered first, followed by partitioning the feature space based on the central rankings. In the third baseline, clustering is applied on a new feature space consisting of both features and label rankings, followed by mapping back to the original feature and ranking space. The RankTree principled approach is based on a Ranking Tree algorithm previously proposed for label ranking prediction. Our modification starts with K random label rankings and iteratively splits the feature space to minimize the ranking loss, followed by re-calculation of the K rankings based on cluster assignments. The MM-PL approach is a multi-prototype supervised clustering algorithm based on the Plackett-Luce (PL) probabilistic ranking model. It represents each cluster with a union of Voronoi cells that are defined by a set of prototypes, and assign each cluster with a set of PL label scores that determine the cluster central ranking. Cluster membership and ranking prediction for a new instance are determined by cluster membership of its nearest prototype. The unknown cluster PL parameters and prototype positions are learned by minimizing the ranking loss, based on two variants of the expectation-maximization algorithm. Evaluation of the proposed algorithms was conducted on synthetic and real-life label ranking data by considering several measures of cluster goodness: (1) cluster compactness in feature space, (2) cluster compactness in label ranking space and (3) label ranking prediction loss. Experimental results demonstrate that the proposed MM-PL and RankTree models are superior to the baseline models. Further, MM-PL is has shown to be much better than other algorithms at handling situations with significant fraction of missing label preferences.