Dynamic Clustering Strategy Research Articles

Fast Progressive Differentiable Architecture Search based on adaptive task granularity reorganization

Shrinkage methods reduce the search space of a Differentiable Architecture Search (DARTS) by progressively discarding candidates, which accelerates the search speed. However, their shrinkage strategy suffers from the vulnerability of too fine task granularity. In other words, they drop only one of the least promising candidates per round of shrinkage, which is suboptimal in terms of performance and efficiency. In this study, we introduce the concept of Granular Computing (GrC) into the shrinkage method and present a Fast Progressive Differentiable Architecture Search (FP-DARTS) method. This method effectively reduces the computational complexity of each round of shrinkage, thereby improving the efficiency and performance of the algorithm. FP-DARTS can be divided into three stages: adaptive granularity division and selection, granular-channel performance evaluation, and progressive shrinkage. In the first stage, to reorganize the task granularity, we cluster the candidate operations into granular-channels and adaptively select the appropriate task granularity. We also propose a dynamic clustering strategy to avoid introducing additional computation. In the second stage, we train the architecture parameters to measure the potential of the granular-channels. In the third stage, to improve the stability of the shrinkage results, we introduce a channel annealing mechanism to smoothly discard unpromising granular-channels. We conducted systematic experiments on CIFAR-10 and ImageNet and achieved a test accuracy of 97.56% on CIFAR-10 with 0.04 GPU-days, and a test accuracy of 75.5% on ImageNet with 1.2 GPU-days. We also conducted experiments on the search space of NAS-Bench-201, and obtained test accuracies of 94.22, 73.07, and 46.23% for CIFAR-10, CIFAR-100 and ImageNet16-120, respectively. The above experimental results demonstrate that FP-DARTS achieves higher search speed and competitive performance compared to other state-of-the-art shrinkage methods and non-shrinkage methods.

CA-CGNet: Component-Aware Capsule Graph Neural Network for Non-Rigid Shape Correspondence

3D non-rigid shape correspondence is significant but challenging in computer graphics, computer vision, and related fields. Although some deep neural networks have achieved encouraging results in shape correspondence, due to the complexity of the local deformation of non-rigid shapes, the ability of these networks to identify the spatial changes of objects is still insufficient. In this paper, we design a Component-aware Capsule Graph Network (CA-CGNet) to further address the features of embedding space based on the component constraints. Specifically, the dynamic clustering strategy is used to classify the features of patches produced by over-segmentation in order to further reduce noise interference. Moreover, aiming at the problem that existing routing ignores the embedding relationship between capsules, we propose a component-aware capsule graph routing to fully describe the relationship between capsules, which regards capsules as nodes in the graph network and constrains nodes through component information. Then, a knowledge distillation strategy is introduced to improve the convergence speed of the network by decreasing the parameters while maintaining accuracy. Finally, a component pair constraint is added to the functional map, and the component-based semantic loss function is proposed, which can compute isomeric in both direct and symmetric directions. The experimental results show that CA-CGNet improves by 10.21% compared with other methods, indicating the accuracy, generalization, and efficiency of our method on the FAUST, SCAPE, TOSCA, and KIDS datasets.

A Two-Stage Strategy for UAV-enabled Wireless Power Transfer in Unknown Environments

Due to the outstanding merits such as mobility, high maneuverability, and flexibility, Unmanned Aerial Vehicles (UAVs) are viable mobile power transmitters that can be rapidly deployed in geographically constrained regions. They are good candidates for supplying power to energy-limited Sensor Nodes (SNs) with Wireless Power Transfer (WPT) technology. In this paper, we investigate a UAV-enabled WPT system that transmits power to a set of SNs at unknown positions. A key challenge is how to efficiently gather the locations of SNs and design a power transfer scheme. We formulate a multi-objective optimization problem to jointly optimize these objectives: maximization of UAV's search efficiency, maximization of total harvested energy, minimization of UAV's flight energy consumption and maximization of UAV's energy utilization efficiency. To tackle these issues, we present a two-stage strategy that includes a UAV Motion Control (UMC) algorithm for obtaining the coordinates of SNs and a Dynamic Genetic Clustering (DGC) algorithm for power transfer via grouping SNs into clusters. First, the UMC algorithm enables the UAV to autonomously control its own motion and conduct target search missions. The objective is to make the energy-restricted UAV find as many SNs as feasible without any apriori knowledge of their information. Second, the DGC algorithm is used to optimize the energy consumption of the UAV by combining a genetic clustering algorithm with a dynamic clustering strategy to maximize the amount of energy harvested by SNs and the energy utilization efficiency of the UAV. Finally, experimental results show that the proposed algorithms outperform their counterparts.

Dynamic clustering based contextual combinatorial multi-armed bandit for online recommendation

Recommender systems still face a trade-off between exploring new items to maximize user satisfaction and exploiting those already interacted with to match user interests. This problem is widely recognized as the exploration/exploitation (EE) dilemma, and the multi-armed bandit (MAB) algorithm has proven to be an effective solution. As the scale of users and items in real-world application scenarios increases, their purchase interactions become sparser. Then three issues need to be investigated when building MAB-based recommender systems. First, large-scale users and sparse interactions increase the difficulty of user preference mining. Second, traditional bandits model items as arms and cannot deal with ever-growing items effectively. Third, widely used Bernoulli-based reward mechanisms only feedback 0 or 1, ignoring rich implicit feedback such as behaviors like click and add-to-cart. To address these problems, we propose an algorithm named Dynamic Clustering based Contextual Combinatorial Multi-Armed Bandits (DC3MAB), which consists of three configurable key components. Specifically, a dynamic user clustering strategy enables different users in the same cluster to cooperate in estimating the expected rewards of arms. A dynamic item partitioning approach based on collaborative filtering significantly reduces the scale of arms and produces a recommendation list instead of one item to provide diversity. In addition, a multi-class reward mechanism based on fine-grained implicit feedback helps better capture user preferences. Extensive empirical experiments on three real-world datasets demonstrate the superiority of our proposed DC3MAB over state-of-the-art bandits (On average, +75.8% in F1 and +54.3% in cumulative reward). The source code is available at https://github.com/HaixHan/DC3MAB.

Towards Optimal Dissemination of Emergency Messages in Internet of Vehicles: A Dynamic Clustering-Based Approach

In this paper, we target dissemination issues of emergency messages in a highly dynamic Internet of Vehicles (IoV) network. IoV is emerging as a new class of vehicular networks to optimize road safety as well as users’ comfort. In such a context, forwarding emergency messages through vehicle-to-vehicle communications (V2V) plays a vital role in enabling road safety-related applications. For instance, when an accident occurs, forwarding such information in real time will help to avoid other accidents in addition to avoiding congestion of network traffic. Thus, dissemination of emergency information is a major concern. However, on the one hand, vehicle density has increased in the last decade which may lead to several issues including message collisions, broadcast storm, and the problem of hidden nodes. On the other hand, high mobility of vehicles and hence dynamic changes of network topology result in failure of dissemination of emergency packets. To overcome these problems, we propose a new dissemination scheme of emergency packets by vehicles equipped with both DSRC and cellular LTE wireless communication capabilities. Our scheme is based on a dynamic clustering strategy, which includes a new cluster head selection algorithm to deal with the broadcast storm problem. Furthermore, our selection algorithm enables not only the election of the most stable vehicles as cluster heads, and hence their exploitation in forwarding the emergency information, but also the avoidance of packet collisions. We simulated our scheme in an urban environment and compared it with other data dissemination schemes. Obtained results show the efficiency of our scheme in minimizing collision and broadcast storm problems, while improving latency, packet delivery ratio and data throughput, as compared to other schemes.

Machine learning based hybrid model for energy efficient secured transmission in wireless sensor networks

Wireless Sensor Networks becomes robust while the nodes possessing the self-configuring capability by which they can be either keeping themselves as a part of the network or they can also simply leave. The transmission of Data is achieved through the inter communication among the nodes participated in the network. For most participative nodes, there is a possibility of energy depletion and thus the network dies despite the other nodes having optimal energy for being operative. Bring decentralized, another classical challenge for WSNs is the threat of intrusion detection which may arise DoS attacks that depleted the energy as well. Major works in the literature focuses either attempts to address the Intrusion detection to fix DoS attacks or keep the optimal energy level by introducing energy preserving strategy and dynamic clustering strategy so that the network lifetime could be extended. This work deals with the hybrid model in which the nodes are clustered together to form the Connected Dominating Set so that the data transmission could be augmented. The objective of the system is to include only nodes with adequate energy that can ensure the validity of the network by marking them as CDS node. The CDS probability of a node depends on foresight of the energy of the participant nodes till the transmission. Packet distribution is done based on the CDS of the node and thus the inappropriate network breakdown that keeps same node waking so often could be avoided. Thereby, retransmission has been avoided due to packet drops, which also conserved the energy of the network. From the simulation results, it is understood that the technique decreases the loss of packets and deferment. The simulation results demonstrated that the proposed model has improved performance metrics such as goodput, throughput, packet dropping ratio. Also, it is seen that the network lifetime has been significantly increased to 50% meantime packet dropping ratio has been reduced to 50% with the energy consumption little lower than 3% when compared with the preliminary enhanced energy-dependent constraint DoS (denial-of-service) detection model and thus the network seems fault-proof and extended reliable.

Clustering-based concurrent topology optimization with macrostructure, components, and materials

This paper proposes a clustering-based concurrent topology optimization method for designing additive manufacturable cellular structures. In the context of multiscale topology optimization, the macrodesign domain can be divided into several subdomains (components) to reduce the number of microstructures which are needed to be optimized. However, in previous works, the division pattern is either artificially decided and fixed during the iteration process or updated according to a fixed parameter range. In this paper, a dynamic clustering strategy is developed to automatically divide the macrodesign domain into several subdomains according to the directions and ratio of the principal stress. K-means algorithm is adopted here for clustering; the advantages are that no predefined range is needed to group the microstructures and the number of microstructures used in the optimization can be arbitrarily specified. The macrostructure and the representative microstructures are optimized simultaneously using a density-based method. Each iteration consists of three sequential steps: First, the macrovariables are updated one step forward. Then, get the stress tensors from the macrovariables updating step and perform the clustering analysis. In the end, update the microvariables one step forward under the current clustering pattern and start the next iteration until the optimization converges. Several numerical examples are presented to demonstrate the effectiveness of the proposed method.

Dynamic Density Clustering Ant Colony Algorithm With Filtering Recommendation Backtracking Mechanism

Ant colony system has a good performance in solving Traveling Salesman Problem (TSP), but it tends to fall into local optimum and is deficient in convergence speed. To address this problem, a dynamic density clustering ant colony algorithm with a filtering recommendation backtracking mechanism is proposed (DBACS). Firstly, a dynamic density clustering strategy is proposed to accelerate the convergence speed of the algorithm and improve the quality of the solution. Under this strategy, the search radius is expanded dynamically to merge adjacent classes, so as to form the differential pheromone distribution. The splicing paths between each class are adjusted through the ant colony algorithm to achieve better performance. Secondly, a recommendation backtracking mechanism based on collaborative filtering is proposed to increase the diversity of the population, thus helping the algorithm jump out of the local optimum. With the help of the collaborative filtering algorithm, some dense data points are recommended for pheromone dynamic backtracking, which can not only help algorithm jump out of the local optimum, but also help the algorithm accelerate convergence. Simulation results show that the improved algorithm can obtain a better solution and higher stability. Especially in solving large-scale TSP, the accuracy of the solution is significantly improved.

Multi-objective particle swarm optimization based on cooperative hybrid strategy

A multi-objective particle swarm optimization based on cooperative hybrid strategy (CHSPSO) is presented in this paper to solve complex multi-objective problems. Most algorithms usually contain only one strategy, which makes them unable to trade off the convergence and diversity when solving the complex multi-objective problems. The proposed cooperative hybrid strategy can effectively guarantee the convergence and the diversity of the algorithm. The multi-population strategy and the dynamic clustering strategy are employed to improve the convergence and the diversity. At the same time, the life strategy and lottery probability selection strategy are used to further ensure the diversity of the population. A series of test functions are used to verify the effectiveness of CHSPSO. The performance of the proposed algorithm is compared with other evolutionary algorithms. The results show that CHSPSO can obtain a better convergence and diversity for the complex multi-objective problems.

The Importance of Being Clustered: Uncluttering the Trends of Statistics from 1970 to 2015

In this paper, we retrace the recent history of statistics by analyzing all the papers published in five prestigious statistical journals since 1970, namely: The Annals of Statistics, Biometrika, Journal of the American Statistical Association, Journal of the Royal Statistical Society, Series B and Statistical Science. The aim is to construct a kind of “taxonomy” of the statistical papers by organizing and clustering them in main themes. In this sense being identified in a cluster means being important enough to be uncluttered in the vast and interconnected world of the statistical research. Since the main statistical research topics naturally born, evolve or die during time, we will also develop a dynamic clustering strategy, where a group in a time period is allowed to migrate or to merge into different groups in the following one. Results show that statistics is a very dynamic and evolving science, stimulated by the rise of new research questions and types of data.

Dynamic User-Centric Clustering for Uplink Cooperation in Multi-Cell Wireless Networks

In this paper, dynamic user-centric cell clustering, which is capable of exploiting dynamics in the channel states, is investigated for joint intra-cluster interference cancellation in a multi-cell environment. A resource efficient cluster size minimization problem is formulated to dynamically group the cells into clusters based on user channel states such that quality-of-service provisioning for cell-edge users can be improved. The integer programming problem depends largely on the intra-cluster interference weight. A subgradient algorithm is employed to solve the relaxed problem when no constrain on cooperation cost is present. To reduce extra burden on backhaul due to base station (BS) cooperation, constraints on the number of per-BS cooperative links and the maximum user-centric cluster size are introduced to the optimization problem, which is solved efficiently by a greedy algorithm. Numerical results show that the proposed dynamic user-centric clustering algorithms achieve significant improvements over existing static and fixed-size dynamic clustering schemes in terms of cell-edge performance and backhaul efficiency. The proposed greedy algorithm, in particular, can effectively alleviate the overall and per-BS cooperation cost while guaranteeing the cooperative gain. With similar resource consumption and outage performance, the proposed scheme achieves 12.6% higher rate gain compared with existing fixed size dynamic clustering strategy.

Satellite observation scheduling with a novel adaptive simulated annealing algorithm and a dynamic task clustering strategy

Abstract Efficient scheduling is significant for the effective use of scarce satellite resources. Task clustering has been demonstrated as an effective strategy for improving the efficiency of satellite scheduling. However, current task clustering strategies are static, i.e., they are integrated into the scheduling in a two-phase manner rather than in a dynamic fashion, without expressing their full potential in improving the satellite scheduling performance. In this study, we present an adaptive simulated annealing–based scheduling algorithm integrated with a dynamic task clustering strategy (ASA-DTC) for satellite observation scheduling problems (SOSPs). Firstly, we develop a formal model for the scheduling of Earth observing satellites. Secondly, we analyse the related constraints involved in the observation task clustering process. Thirdly, we detail an implementation of the dynamic task clustering strategy and the adaptive simulated annealing algorithm. The adaptive simulated annealing algorithm is efficient and contains sophisticated mechanisms, i.e., adaptive temperature control, tabu-list-based short-term revisiting avoidance mechanism and intelligent combination of neighbourhood structures. Finally, we report on experimental simulation studies to demonstrate the competitive performance of ASA-DTC. We show that ASA-DTC is particularly effective when SOSPs contain a large number of targets or when these targets are densely distributed in a certain area.

An Improved Brain Storm Optimization with Dynamic Clustering Strategy

Intelligence algorithms play an increasingly important role in the field of intelligent control. Brain storm optimization (BSO) is a new kind of swarm intelligence algorithm inspired by emulating the collective behavior of human beings in the problem solving process. To improve the performance of the original BSO, many variants of BSO are proposed. In this paper, an improved BSO algorithm with dynamic clustering strategy (BSO-DCS) is proposed as a variant of BSO for global optimization problems. The basic framework of BSO is firstly introduced. Then to reduce the time complexity of the original BSO, a new grouping method named dynamic clustering strategy (DCS) is proposed to improve the clustering method in the original BSO. To verify the effectiveness of the proposed BSO-DCS, it is tested on 12 benchmark functions of CEC 2005 with 30 dimensions. Experimental results show that DCS is an effective strategy to reduce the time complexity, and the improved BSO-DCS performs greatly better than the original BSO algorithm.

Dynamic and Fault-Tolerant Clustering for Scientific Workflows

Task clustering has proven to be an effective method to reduce execution overhead and to improve the computational granularity of scientific workflow tasks executing on distributed resources. However, a job composed of multiple tasks may have a higher risk of suffering from failures than a single task job. In this paper, we conduct a theoretical analysis of the impact of transient failures on the runtime performance of scientific workflow executions. We propose a general task failure modeling framework that uses a maximum likelihood estimation-based parameter estimation process to model workflow performance. We further propose three fault-tolerant clustering strategies to improve the runtime performance of workflow executions in faulty execution environments. Experimental results show that failures can have significant impact on executions where task clustering policies are not fault-tolerant, and that our solutions yield makespan improvements in such scenarios. In addition, we propose a dynamic task clustering strategy to optimize the workflow's makespan by dynamically adjusting the clustering granularity when failures arise. A trace-based simulation of five real workflows shows that our dynamic method is able to adapt to unexpected behaviors, and yields better makespans when compared to static methods.

A dynamic clustering strategy in a demand paging environment

An algorithm is presented which dynamically clusters pages of a problem program based on its past program behavior (i.e. reference string patterns) in a demand paged virtual memory environment. The objective of this algorithm is to minimize the number of page faults during execution, while at the same time use memory page frames efficiently. Dynamic clusters of time and reference related pages are built during a program's execution time. Whenever a page fault for the i-th page occurs in this time evolving environment, the pages of the cluster associated with the i-th page is compared to the pages currently in real (physical) memory. Thus during this current page fault, the demanded page, and any associated clustered pages not currently in physical memory are placed into memory. Page frames holding pages not in the current cluster are returned to the memory management system. Thus the physical amount of memory allocated to a processing program is dependent upon the size of the associated cluster at that time . Simulation results of program behavior operating under this dynamic clustering strategy indicates that significant improvements in page faults and in the space time product may be achieved. The algorithm requires fewer real page frames per executed instruction than most currently implemented algorithms that utilize a fixed number of page frames per problem programs (i.e. fixed allocated partition or region). The algorithm, MLMM (Modified Locality Matrix Model), an extension of work by Hedges and Pooch [12] is used to determine inherent program locality to predict independent dynamic program behavior, separating instruction from data references. Furthermore, strength coefficients between weakly or loosely coupled clusters are used to refine the cluster population, identify cluster transitions, as well as indicate the behavior of the cluster formations.