Efficient frameworks for greedy split and new depth first search split procedures for routing problems

Split procedures have proven their efficiency within global optimization frameworks for routing problems by splitting giant tours into trips. This is done by generating an optimal shortest path within an auxiliary graph constructed from the giant tour. This article provides a state of the art of split practice in routing problems and gives its key features. The efficiency of the method critically depends on the node-splitting procedure and on the upper and lower bound approximations. Suitable complexity can be obtained using a new depth first search procedure which is introduced here and provides a new algorithm that is specially designed for large scale problems with resource constraints. Experiments show that the depth first search split procedure introduced in this article, used as an evaluation function in a global framework, can improve the results obtained by the use of the “classical” split procedure on the Location-Routing Problem and the Heterogeneous Vehicle Routing Problem. The Location-Routing Problem is also used to provide further analysis and a fair comparative study between the two split versions.     

A CNN-based personalized system for attention detection in wayfinding tasks

Firefighters are often exposed to extensive wayfinding information in various formats owing to the increasing complexity of the built environment. Because of the individual differences in processing assorted types of information, a personalized cognition-driven intelligent system is necessary to reduce the cognitive load and improve the performance in the wayfinding tasks. However, the mixed and multi-dimensional information during the wayfinding tasks bring severe challenges to intelligent systems in detecting and nowcasting the attention of users. In this research, a virtual wayfinding experiment is designed to simulate the human response when subjects are memorizing or recalling different wayfinding information. Convolutional neural networks (CNNs) are designed for automated attention detection based on the power spectrum density of electroencephalography (EEG) data collected during the experiment. The performance of the personalized model and the generalized model are compared and the result shows a personalized CNN is a powerful classifier in detecting the attention of users with high accuracy and efficiency. The study thus will serve a foundation to support the future development of personalized cognition-driven intelligent systems.     

SAG-Net：用于联合视盘和视杯分割的新型跳过注意力指导网络

学习特征图语义信息和位置信息对于在视网膜图像分割中产生理想的结果至关重要。最近,卷积神经网络在提取特征图有效信息方面已经表现出了强大的能力,然而,卷积和池化操作会过滤掉一些有用的信息。提出了一种新型跳过注意力指导网络SAG-Net来保存特征图语义和位置信息并指导扩展工作。在SAG-Net中,首先引入了跳过注意力门SAtt模块,将其用作敏感的扩展路径来传递先前特征图的语义信息和位置信息,不仅有助于消除噪声,还进一步减小了背景的负面影响。其次,通过合并图像金字塔保留上下文特征来进一步优化SAG-Net。在Drishti-GS1数据集上,联合视盘和视杯分割任务表明了SAG-Net的有效性。综合结果表明,SAG-Net优于原始的U-Net方法以及其他用于视盘和视杯分割的最新的方法。     

A large neighbourhood search heuristic for ship routing and scheduling with split loads

The purpose of this paper is to present and solve a new, important planning problem faced by many shipping companies dealing with the transport of bulk products. These shipping companies are committed to carrying some contract cargoes and will try to derive additional revenue from optional spot cargoes. In most of the literature on ship routing and scheduling problems a cargo cannot be transported by more than one ship. By introducing split loads this restriction is removed and each cargo can be transported by several ships. In this paper we propose a large neighbourhood search heuristic for the ship routing and scheduling problem with split loads. Computational results show that the heuristic provides good solutions to real-life instances within reasonable time. It is also shown that introducing split loads can yield significant improvements.     

Designing for user attention: A method for supporting unobtrusive routine tasks

The automation of user routine tasks is one of the most important challenges in the development of Ambient Intelligence systems. However, this automation may be annoying since some tasks may grab users attention in inappropriate situations. Since user attention is a valuable resource, task automation must behave in a considerate manner demanding user attention only when it is required. To address this issue, this work presents a systematic method for supporting the design and automation of unobtrusive routine tasks that can adjust their obtrusiveness level at runtime according to the user attentional resources and context. This method proposes to design the routine tasks that the system must carry out and how they must interact with users in terms of obtrusiveness. The method also provides a software infrastructure that makes the execution of the tasks at the appropriate obtrusiveness degree a reality. Finally, the system has been validated by means of usefulness and performance tests and a practical case study that demonstrates the correctness and applicability of our approach without compromising system performance.     

A branch-price-and-cut method for a ship routing and scheduling problem with split loads

We present a branch-price-and-cut method to solve a maritime pickup and delivery problem with time windows and split loads. The fleet of ships is heterogeneous and fixed for the planning horizon, and no common depot exists. The cargoes picked up and delivered consist of both mandatory and optional cargoes, and each cargo may be split among several ships. The objective is to design a route for each ship that will maximize the total profit from transporting all the mandatory and a subset of the optional cargoes. To solve this problem we introduce a new path-flow formulation, which we solve by branch-price-and-cut. The subproblem is a new variant of the elementary shortest path problem with resource constraints, where a multi-dimensional knapsack problem is solved to compute optimal cargo quantities. Further, we present new valid inequalities for this problem, and adaptations of existing inequalities successfully used to solve related problems in the literature. Finally, the computational results show that for certain types of instances, our solution method outperforms existing methods proposed in the literature.     

Optimizing the stretch of independent tasks on a cluster: From sequential tasks to moldable tasks

This paper addresses the problem of scheduling non-preemptive moldable tasks to minimize the stretch of the tasks in an online non-clairvoyant setting. To the best of the authors’ knowledge, this problem has never been studied before. To tackle this problem, first the sequential subproblem is studied through the lens of the approximation theory. An algorithm, called DASEDF, is proposed and, through simulations, it is shown to outperform the first-come, first-served scheme. Furthermore, it is observed that machine availability is the key to getting good stretch values. Then, the moldable task scheduling problem is considered, and, by leveraging the results from the sequential case, another algorithm, DBOS, is proposed to optimize the stretch while scheduling moldable tasks. This work is motivated by a task scheduling problem in the context of parallel short sequence mapping which has important applications in biology and genetics. The proposed DBOS algorithm is evaluated both on synthetic data sets that represent short sequence mapping requests and on data sets generated using log files of real production clusters. The results show that the DBOS algorithm significantly outperforms the two state-of-the-art task scheduling algorithms on stretch optimization.     

Learner-controlled selection of tasks with different surface and structural features: Effects on transfer and efficiency

Surface task features are more salient than structural task features and thus easier to recognize for novices. It is predicted that the more salient the task features the better learners can choose personally relevant and varied tasks, which enhances learning transfer. To investigate this prediction, a 2 × 2 factorial experiment with 72 participants studied the effects of control over tasks that differ in their surface features (learner, program) and in their structural features (learner, program). Learner control over the selection of tasks with salient surface features enables learners to select personally relevant and varied tasks. This is believed to yield higher effectiveness (i.e., higher near and far transfer test performance) as well as higher efficiency (i.e., higher transfer test performance combined with lower associated mental effort). Learner control over the selection of tasks with non-salient structural features does not enable learners to select personally relevant and varied tasks and is therefore not expected to yield beneficial effects on learning. The results show positive effects of learner control over the selection of tasks with salient surface features for efficiency on the far transfer test but not for effectiveness. Theoretical and practical implications are discussed.     

The multi-depot split delivery vehicle routing problem: An integer programming-based heuristic, new test problems, and computational results

The multi-depot split delivery vehicle routing problem combines the split delivery vehicle routing problem and the multiple depot vehicle routing problem. We define this new problem and develop an integer programming-based heuristic for it. We apply our heuristic to 30 instances to determine the reduction in distance traveled that can be achieved by allowing split deliveries among vehicles based at the same depot and vehicles based at different depots. We generate new test instances with high-quality, visually estimated solutions and report results on these instances.     

LSF参数转换分裂矢量量化的卡尔曼后滤波增强方法

针对LSP参数在转换分裂矢量量化（Switch Split Vector Quantization,SSVQ）中未能充分利用子矢量间相关性的不足,提出了一种LSP参数SSVQ的卡尔曼后滤波增强方法。该方法在解码端利用卡尔曼滤波器来进一步发掘子矢量间和连续帧矢量间的相关性,并结合SSVQ中分类的转换来自适应地调整卡尔曼滤波器的参数。实验结果表明,方法可在SSVQ的平均频谱误差为0.9~1.0 dB时进一步减少0.01~0.02 dB。     

PMA-Net: A parallelly mixed attention network for person re-identification

The objective of person re-identification (Re-ID) is to match unmistakable people across different settings and camera views. Although the use of convolutional neural networks (CNN) has been effective, they lose much significant data and ignore the correlation between the overall and partial features brought about by downsampling operators (e.g., pooling). This compromises the retrieval performance based on the Euclidean distance. We propose a parallelly mixed attention network (PMA-Net) to demonstrate the complementarity of spatial and channel information for Re-ID. First, we integrate self-attention with depthwise convolution in a parallel design, extracting different dimensions of information. Second, we propose two interactive branches to entwine spatial and channel information while extracting features. Then, we add a new additional embedding to incorporate non-visual inputs and reduce the feature bias toward camera variances. Finally, we perform additional maximizing the minimal pairwise angles (MMA) regularization on the features of PMA, which successfully encourages the angular diversity of feature vectors. We confirm the effectiveness of PMA-Net through an extensive set of ablation studies. The results of three benchmarks show the superiority of our model over current mAP/top1 accuracies of 91.3%/96.1% on Market-1501, 83.3%/91.7% on DukeMTMC-ReID, and 66.2%/85.1% on MSMT17.     

A distributed self-scheduler for partially ordered tasks

Large scale scientific applications such as weather modelling and continuous simulation require the orders of magnitude performance improvement available with the new generation of parallel vector supercomputers such as the Floating Point Systems T Series. Many of these applications exhibit a high degree of parallelism, much of which can be expressed as computational tasks which are of varying size and degrees of dependence on one another, and can be partially ordered for execution. DeSPOT (A Distributed Self-Scheduler for Partially Ordered Tasks) is an algorithm for the dynamic distribution of such non-uniform tasks to achieve automatic load balancing on a distributed memory hypercube multiprocessor. This paper describes the DeSPOt algorithm and presents its performance characteristics of various test cases using result timings on the FPS T 20.     

Acquisition of troubleshooting skills in a computer simulation: Worked example vs. conventional problem solving instructional strategies

In a computer-based simulation of a chemical processing plant, the differential effects of three instructional strategies for learning how to troubleshoot the plant’s malfunctions were investigated. In an experiment concerning learners’ transfer performance and mental effort, the simulation presented the three strategies to three groups of learners and measured their performance on the transfer tasks. In this experiment, conventional problem solving was contrasted with two worked example strategies. The results indicated a significant difference between practicing problem solving and using worked examples. Learners who practiced problem solving in an interactive simulation outperformed the learners who studied computer-based worked examples. They also invested lower mental effort in transfer tasks. When accounting for the difference in the learners’ domain knowledge, the strategies were not significantly different among the more experienced learners. For the less experienced learners, those who practiced problem solving significantly outperformed their worked example counterparts. Among all participants and also among less experienced learners the problem solving group invested significantly lower mental effort in the performance of transfer tasks. Based on the results of this study, the authors recommend the use of the conventional problem solving strategy with or without worked examples for learning complex skills.     

A visual attention model for robot object tracking

Inspired by human behaviors, a robot object tracking model is proposed on the basis of visual attention mechanism, which is fit for the theory of topological perception. The model integrates the image-driven, bottom-up attention and the object-driven, top-down attention, whereas the previous attention model has mostly focused on either the bottom-up or top-down attention. By the bottom-up component, the whole scene is segmented into the ground region and the salient regions. Guided by top-down strategy which is achieved by a topological graph, the object regions are separated from the salient regions. The salient regions except the object regions are the barrier regions. In order to estimate the model, a mobile robot platform is developed, on which some experiments are implemented. The experimental results indicate that processing an image with a resolution of 752*480 pixels takes less than 200ms and the object regions are unabridged. The analysis obtained by comparing the proposed model with the existing model demonstrates that the proposed model has some advantages in robot object tracking in terms of speed and efficiency.     

Modeling and evaluation of attention allocation in an intelligent tractor head‐up display interface

Tractors are becoming increasingly intelligent due to developments in electronics and information technology. Amid the large quantities of information in the driving cab, drivers may lack the attention resources demanded when operating, thus leading to neglect of important information. Therefore, the interface design stage must consider the human factors of the design. Designers must study and optimize the factors influencing attention allocation related to the interface design. In this study, the criticality of information is first proposed for quantifying the importance of various information in the display interface. Subsequently, the factors affecting attention allocation are investigated through experimental methods. Finally, a matching index is proposed to evaluate the interface design with respect to attention allocation. This enables the optimization of the interface design by matching attention allocation and criticality of the information. The matching index can thus serve as a design specification for tractor head‐up display interfaces.     

A KinFu based approach for robot spatial attention and view planning

When a user and a robot share the same physical workspace the robot may need to keep an updated 3D representation of the environment. Indeed, robot systems often need to reconstruct relevant parts of the environment where the user executes manipulation tasks. This paper proposes a spatial attention approach for a robot manipulator with an eye-in-hand Kinect range sensor. Salient regions of the environment, where user manipulation actions are more likely to have occurred, are detected by applying a clustering algorithm based on Gaussian Mixture Models applied to the user hand trajectory. A motion capture sensor is used for hand tracking. The robot attentional behavior is driven by a next-best view algorithm that computes the most promising range sensor viewpoints to observe the detected salient regions, where potential changes in the environment have occurred. The environment representation is built upon the PCL KinFu Large Scale project [1], an open source implementation of KinectFusion. KinFu has been modified to support the execution of the next-best view algorithm directly on the GPU and to properly manage voxel data. Experiments are reported to illustrate the proposed attention based approach and to show the effectiveness of GPU-based next-best view planning compared to the same algorithm executed on the CPU.     

A new method for solving kinematic tasks for robots

The main difficulties in solving inverse kinematic tasks for robots are singularities and nonlinearities. In the first part of this paper practical conditions regarding singularities of 6 df wrist-joint structure robots are presented. These considerations are based on a simple iterative inversion method by the use of which introduction and characterization of the concepts of “essential” and “procedural” singularities are realized. The second part of the paper deals with the question of nonlinearities. It is pointed out that in robotics there are two different kinds of nonlinearities. The first type is present because the sine and cosine functions of the joint coordinates appear in the formulae of direct kinematic tasks. The second type is of finite polynomial nature, and originates from the quadratic functions of the unit vectors parallel to the rotation axes of the joints of the arm. The authors point out that an appropriate quaternion representation of the robot kinematics offers a convenient possibility to get rid of sine- and cosine-type nonlinearities. Those nonlinearities which remain are the results of simple quadratic operations (mainly multiplications and additions, subtractions and divisions); therefore the proposed method seems to be especially appropriate for realizations based on neural networks or particular hardwares using RISC processors. The advantages and disadvantages of this new method are compared to those of another solution recently published by the authors.     

A collaborative gated attention network for fine-grained visual classification

Fine-grained image classification aims at subdividing large coarse-grained categories into finer-grained subcategories. Most existing fine-grained research methods use a single attention mechanism or multiple sub-networks to zoom in and find distinguishable local feature regions. These models seldom explore the intrinsic connections between cross-layer features with similar semantic features. This tends to show erratic performance in images with complex backgrounds. To this end, we propose a feature-semantic fusion module to enhance the diversity of global feature information. Second, we employ cross-layer spatial attention and channel attention modules, which can accurately locate local key regions of images. Finally, we propose a cross-gate attention module that can find rich discriminative features from key object regions of images to guide the final classification. Experiments show that the proposed model performs well on three datasets: CUB-200-2011, Stanford cars, and FGVC aircraft.     

A fusion spatial attention approach for few-shot learning

Few-shot learning is a challenging problem in computer vision that aims to learn a new visual concept from very limited data. A core issue is that there is a large amount of uncertainty introduced by the small training set. For example, the few images may include cluttered backgrounds or different scales of objects. Existing approaches mostly address this problem from either the original image space or the embedding space by using meta-learning. To the best of our knowledge, none of them tackle this problem from both spaces jointly. To this end, we propose a fusion spatial attention approach that performs spatial attention in both image and embedding spaces. In the image space, we employ a Saliency Object Detection (SOD) module to extract the saliency map of an image and provide it to the network as an additional channel. In the embedding space, we propose an Adaptive Pooling (Ada-P) module tailored to few-shot learning that introduces a meta-learner to adaptively fuse local features of the feature maps for each individual embedding. The fusion process assigns different pooling weights to the features at different spatial locations. Then, weighted pooling can be conducted over an embedding to fuse local information, which can avoid losing useful information by considering the spatial importance of the features. The SOD and Ada-P modules can be used within a plug-and-play module and incorporated into various existing few-shot learning approaches. We empirically demonstrate that designing spatial attention methods for few-shot learning is a nontrivial task and our method has proven effective for it. We evaluate our method using both shallow and deeper networks on three widely used few-shot learning benchmarks, miniImageNet, tieredImageNet and CUB, and demonstrate very competitive performance.     

Multi-input single-output control for extending the operating range: Generalized split range control using the baton strategy

Split range control is used to extend the steady-state operating range for a single output (controlled variable) by using multiple inputs (manipulated variables). The standard implementation of split range control uses a single controller with a split range block, but this approach has limitations when it comes to tuning. In this paper, we introduce a generalized split range control structure that overcomes these limitations by using multiple independent controllers with the same setpoint. Undesired switching between the controllers is avoided by using a baton strategy where only one controller is active at a time. As an alternative solution we consider model predictive control (MPC), but it requires a detailed dynamic model and does not allow for using only one input at a time.