Pasar al contenido principal

Resultados relacionados

Resultados totales (Incluyendo duplicados): 16
Encontrada(s) 1 página(s)

RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Cloud Workflow Scheduling with Deadlines and Time Slot Availability

  • Li, Xiaoping
  • Qian, L.
  • Ruiz García, Rubén
[EN] Allocating service capacities in cloud computing is based on the assumption that they are unlimited and can be used at any time. However, available service capacities change with workload and cannot satisfy users' requests at any time from the cloud provider's perspective because cloud services can be shared by multiple tasks. Cloud service providers provide available time slots for new user's requests based on available capacities. In this paper, we consider workflow scheduling with deadline and time slot availability in cloud computing. An iterated heuristic framework is presented for the problem under study which mainly consists of initial solution construction, improvement, and perturbation. Three initial solution construction strategies, two greedy-and fair-based improvement strategies and a perturbation strategy are proposed. Different strategies in the three phases result in several heuristics. Experimental results show that different initial solution and improvement strategies have different effects on solution qualities., This work has been supported by the National Natural Science Foundation of China (Nos. 61572127, 61272377) and the Key Research & Development Program in Jiangsu Province (No. BE2015728). Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project "RESULT - Realistic Extended Scheduling Using Light Techniques" (No. DPI2012-36243-C02-01) partially financed with FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

An iterated greedy heuristic for no-wait flow shops with sequence dependent setup times, learning and forgetting effects

  • Li, Xiaoping
  • Yang, Z.
  • Ruiz García, Rubén
  • Chen, T.
  • Sui, S.
[EN] This paper addresses a sequence dependent setup times no-wait flowshop with learning and forgetting effects to minimize total flowtime. This problem is NP-hard and has never been considered before. A position-based learning and forgetting effects model is constructed. Processing times of operations change with the positions of corresponding jobs in a schedule. Objective increment properties are deduced and based on them three accelerated neighbourhood construction heuristics are presented. Because of the simplicity and excellent performance shown in flowshop scheduling problems, an iterated greedy heuristic is proposed. The proposed iterated greedy algorithm is compared with some existing algorithms for related problems on benchmark instances. Comprehensive computational and statistical tests show that the presented method obtains the best performance among the compared methods. (C) 2018 Elsevier Inc. All rights reserved., This work is supported by the National Natural Science Foundation of China (Nos. 61572127, 61272377), the Collaborative Innovation Center of Wireless Communications Technology and the Key Natural Science Fund for Colleges and Universities in Jiangsu Province (No. 12KJA630001). Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness(MINECO), under the project "SCHEYARD - Optimization of Scheduling Problems in Container Yards" with reference DPI2015-65895-R.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Group Scheduling With Nonperiodical Maintenance and Deteriorating Effects

  • Xu, Haiyan
  • Li, Xiaoping
  • Ruiz García, Rubén
  • Zhu, Haihong
[EN] In this paper, we consider single-machine group scheduling with nonperiodical maintenance and deteriorating effects. Nonperiodical maintenance, which has unfixed maintaining interval or the number of jobs in each group is unfixed, results in a variable number of groups. Deteriorating effects lead to longer processing times of which the deterioration index depends on job grouping. This problem is of significance in different production settings and is much more difficult than and general that other simpler single-machine group scheduling problems. Making use of historical processing times, we construct the actual processing time model for jobs. We prove that the problem under study is NP-hard. By transforming the optimization objective, properties are discovered and two batch-based heuristics are presented for small size problems. To further improve the effectiveness for large size problems, an iterated greedy algorithm is proposed being its main advantages simplicity and effectiveness. The proposed methods are evaluated over a large number of random instances with calibrated parameters and components. Comprehensive computational and statistical analyses demonstrate the superiority of the methods proposed over adapted existing approaches, This work was supported in part by the National Key Research and Development Program of China under Grant 2017YFB1400801, in part by the National Natural Science Foundation of China under Grant 61572127, Grant 61872077, and Grant 61832004, and in part by the Collaborative Innovation Center of Wireless Communications Technology. The work of R. Ruiz was supported by the Spanish Ministry of Economy and Competitiveness through the Project "SCHEYARD-Optimization of Scheduling Problems in Container Yards" with FEDER funds under Grant DPI2015-65895-R. This paper was recommended by Associate Editor W. Shen
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Hybrid resource provisioning for cloud workflows with malleable and rigid tasks

  • Chen, Long
  • Li, Xiaoping
  • Guo, Yucheng
  • Ruiz García, Rubén
[EN] In cloud computing, reserved and on-demand instances are generally provided by service providers. Hybridization of the two alternatives can considerably save costs when renting resources from the cloud. However, it is a big challenge to determine the appropriate amount of reserved and on-demand resources in terms of users' requirements. In this paper, the workflow scheduling problem with both reserved and on-demand instances is considered. The objective is to minimize the total rental cost under deadline constrains. The considered problem is mathematically modeled. A multiple sequence-based earliest finish time method is proposed to construct schedules for the workflows. Four different rules are used to generate initial task allocation sequences. Types and quantities of resources are determined by a free time block-based schedule construction mechanism. New sequences are generated by a variable neighborhood search method. Experimental and statistical analyses and results demonstrate that the proposed algorithm algorithm generates considerable cost savings when compared to the algorithms with only on-demand or reserved instances., l This work is supported by the National Key Research and Development Program of China (No. 2017YFB1400801), the National Natural Science Foundation of China (Nos. 61572127, 61872077, 61832004) and Collaborative Innovation Center of Wireless Communications Technology. Rub~en Ruiz is supported by the Spanish Ministry of Economy and Competitiveness, under the project "SCHEYARD-Optimization of scheduling problems in container yards" (No. DPI2015-65895-R) partly financed with FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Weighted General Group Lasso for Gene Selection in Cancer Classification

  • Wang, Yadi
  • Li, Xiaoping
  • Ruiz García, Rubén
[EN] Relevant gene selection is crucial for analyzing cancer gene expression datasets including two types of tumors in cancer classification. Intrinsic interactions among selected genes cannot be fully identified by most existing gene selection methods. In this paper, we propose a weighted general group lasso (WGGL) model to select cancer genes in groups. A gene grouping heuristic method is presented based on weighted gene co-expression network analysis. To determine the importance of genes and groups, a method for calculating gene and group weights is presented in terms of joint mutual information. To implement the complex calculation process of WGGL, a gene selection algorithm is developed. Experimental results on both random and three cancer gene expression datasets demonstrate that the proposed model achieves better classification performance than two existing state-of-the-art gene selection methods., This work was supported in part by the National Natural Science Foundation of China under Grant 61572127, in part by the National Key Research and Development Program of China under Grant 2017YFB1400801, in part by the Key Research and Development Program in Jiangsu Province under Grant BE2015728, and in part by the Collaborative Innovation Center of Wireless Communications Technology. The work of R. Ruiz was supported by the Spanish Ministry of Economy and Competitiveness through the Project "SCHEYARD-Optimization of Scheduling Problems in Container Yards" partly financed with FEDER funds under Grant DPI2015-65895-R. This paper was recommended by Associate Editor S. Yang.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Resource Renting for Periodical Cloud Workflow Applications

  • Chen, Long
  • Li, Xiaoping
  • Ruiz García, Rubén
[EN] Cloud computing is a new resource provisioning mechanism, which represents a convenient way for users to access different computing resources. Periodical workflow applications commonly exist in scientific and business analysis, among many other fields. One of the most challenging problems is to determine the right amount of resources for multiple periodical workflow applications. In this paper, the periodical workflow applications scheduling problem with total renting cost minimization is considered. The novelty of this work relies precisely on this objective function, which is more realistic in practice than the more commonly considered makespan minimization. An integer programming model is constructed for the problem under study. A Precedence Tree based Heuristic (PTH) is developed which considers three types of initial schedule construction methods. Based on the initial schedule, two improvement procedures are presented. The proposed methods are compared with existing algorithms for the related makespan based multiple workflow scheduling problem. Experimental and statistical results demonstrate the effectiveness and efficiency of the proposed algorithm., This work is supported by the National Natural Science Foundation of China (No. 61572127, 61272377), the Key Research & Development program in Jiangsu Province (No. BE2015728) and Collaborative Innovation Center of Wireless Communications Technology. Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project "SCHEYARD-Optimization of Scheduling Problems in Container Yards" (No. DPI2015-65895-R) financed by FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

An Exact Algorithm for the Shortest Path Problem With Position-Based Learning Effects

  • Wang, Yamin
  • Li, Xiaoping
  • Ruiz García, Rubén
[EN] The shortest path problems (SPPs) with learning effects (SPLEs) have many potential and interesting applications. However, at the same time they are very complex and have not been studied much in the literature. In this paper, we show that learning effects make SPLEs completely different from SPPs. An adapted A* (AA*) is proposed for the SPLE problem under study. Though global optimality implies local optimality in SPPs, it is not the case for SPLEs. As all subpaths of potential shortest solution paths need to be stored during the search process, a search graph is adopted by AA* rather than a search tree used by A*. Admissibility of AA* is proven. Monotonicity and consistency of the heuristic functions of AA* are redefined and the corresponding properties are analyzed. Consistency/monotonicity relationships between the heuristic functions of AA* and those of A* are explored. Their impacts on efficiency of searching procedures are theoretically analyzed and experimentally evaluated., This work was supported in part by the National Natural Science Foundation of China under Grant 61572127 and Grant 61272377, and in part by the Specialized Research Fund for the Doctoral Program of Higher Education under Grant 20120092110027. The work of R. Ruiz was supported in part by the Spanish Ministry of Economy and Competitiveness under Project "RESULT-Realistic Extended Scheduling Using Light Techniques" under Grant DPI2012-36243-C02-01, and in part by the FEDER.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Methods for Scheduling Problems Considering Experience, Learning, and Forgetting Effects

  • Li, Xiaoping
  • Jiang, Y.
  • Ruiz García, Rubén
[EN] Workers with different levels of experience and knowledge have different effects on job processing times. By taking into account 1) the sum-of-processing-time; 2) the job-position; and 3) the experience of workers, a more general learning model is introduced for scheduling problems. We show that this model generalizes existing ones and brings the consideration of learning and forgetting effects closer to reality. We demonstrate that some single machine scheduling problems are polynomially solvable under this general model. Considering the forgetting effect caused by the idle time on the second machine, we construct a learning-forgetting model for the two-machine permutation flow shop scheduling problem with makespan minimization. A branch-and-bound method and four heuristics are presented to find optimal and approximate solutions, respectively. The proposed heuristics are evaluated over a large number of randomly generated instances. Experimental results show that the proposed heuristics are effective and efficient., This work was supported in part by the National Natural Science Foundation of China under Grant 61572127 and Grant 61272377, in part by the Key Research and Development Program in Jiangsu Province under Grant BE2015728, in part by the Collaborative Innovation Center of Wireless Communications Technology and the Key Natural Science Fund for Colleges and Universities in Jiangsu Province under Grant 12KJA630001, and in part by the Collaborative Innovation Center of Wireless Communications Technology. The work of R. Ruiz was supported by the Spanish Ministry of Economy and Competitiveness through Project "SCHEYARD-Optimization of Scheduling Problems in Container Yards" under Grant DPI2015-65895-R. This paper was recommended by Associate Editor A. Janiak.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

A delay-based dynamic scheduling algorithm for bag-of-task workflows with stochastic task execution times in clouds

  • Cai, Zhicheng
  • Li, Xiaoping
  • Ruiz García, Rubén
  • Li, Qianmu
[EN] Bag-of-Tasks (BoT) workflows are widespread in many big data analysis fields. However, there are very few cloud resource provisioning and scheduling algorithms tailored for BoT workflows. Furthermore, existing algorithms fail to consider the stochastic task execution times of BoT workflows which leads to deadline violations and increased resource renting costs. In this paper, we propose a dynamic cloud resource provisioning and scheduling algorithm which aims to fulfill the workflow deadline by using the sum of task execution time expectation and standard deviation to estimate real task execution times. A bag-based delay scheduling strategy and a single-type based virtual machine interval renting method are presented to decrease the resource renting cost. The proposed algorithm is evaluated using a cloud simulator ElasticSim which is extended from CloudSim. The results show that the dynamic algorithm decreases the resource renting cost while guaranteeing the workflow deadline compared to the existing algorithms. (C) 2017 Elsevier B.V. All rights reserved., The authors would like to thank the reviewers for their constructive and useful comments. This work is supported by the National Natural Science Foundation of China (Grant No. 61602243 and 61572127), the Natural Science Foundation ofJiangsu Province (Grant No. BK20160846), Jiangsu Key Laboratory of Image and Video Understanding for Social Safety (Nanjing University of Science and Technology, Grant No. 30916014107), the Fundamental Research Funds for the Central University (Grant No. 30916015104). Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project "SCHEYARD" (No. DP12015-65895-R) co-financed by FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Scheduling Stochastic Multi-Stage Jobs to Elastic Hybrid Cloud Resources

  • Zhu, Jie
  • Li, Xiaoping
  • Ruiz García, Rubén
  • Xu, Xiaolong
[EN] We consider a special workflow scheduling problem in a hybrid-cloud-based workflow management system in which tasks are linearly dependent, compute-intensive, stochastic, deadline-constrained and executed on elastic and distributed cloud resources. This kind of problems closely resemble many real-time and workflow-based applications. Three optimization objectives are explored: number, usage time and utilization of rented VMs. An iterated heuristic framework is presented to schedule jobs event by event which mainly consists of job collecting and event scheduling. Two job collecting strategies are proposed and two timetabling methods are developed. The proposed methods are calibrated through detailed designs of experiments and sound statistical techniques. With the calibrated components and parameters, the proposed algorithm is compared to existing methods for related problems. Experimental results show that the proposal is robust and effective for the problems under study., This work is sponsored by the National Natural Science Foundations of China (Nos. 71401079, 61572127, 61472192), the National Key Research and Development Program of China (No. 2017YFB1400801) and the Collaborative Innovation Center of Wireless Communications Technology. Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project "SCHEYARD-Optimization of Scheduling Problems in Container Yards" (No. DPI2015-65895-R) financed by FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Idle block based methods for cloud workflow scheduling with preemptive and non-preemptive tasks

  • Chen, Long
  • Li, Xiaoping
  • Ruiz García, Rubén
[EN] Complex workflow applications are widely used in scientific computing and economic analysis, which commonly include both preemptive and non-preemptive tasks. Cloud computing provides a convenient way for users to access different resources based on the ¿pay-as-you-go¿ model. However, different resource renting alternatives (reserved, on-demand or spot) are usually provided by the service provider. The spot instances provide a dynamic and cheaper alternative comparing to the on-demand one. However, failures often occur due to the fluctuations of the price of the instance. It is a big challenge to determine the appropriate amount of spot and on-demand resources for workflow applications with both preemptive and non-preemptive tasks. In this paper, the workflow scheduling problem with both spot and on-demand instances is considered. The objective is to minimize the total renting cost under deadline constrains. An idle time block-based method is proposed for the considered problem. Different idle time block-based searing and improving strategies are developed to construct schedules for workflow applications. Schedules are improved by a forward and backward moving mechanism. Experimental and statistical results demonstrate the effectiveness of the proposed algorithm over a lot of tests with different sizes., This work is supported by the National Natural Science Foundation of China (No. 61572127, 61272377), the National Key Research and Development Program of China (No. 2017YFB1400800). Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project "SCHEYARD - Optimization of Scheduling Problems in Container Yards" (No. DPI2015-65895-R) financed by FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

An effective heuristic for project scheduling with resource availability cost

  • Zhu, Xia
  • Ruiz García, Rubén
  • Li, Shiyu
  • Li, Xiaoping
[EN] The resource constrained project scheduling problem (RCPSP) is widely studied in the literature and has a host of applications in practice. As a variant of the RCPSP, the resource availability cost problem (RACP), which has the aim of minimizing the availability costs of renewable resources in order to complete a project subject to a given deadline, is considered in this paper. We divide the RACP into two sub-problems: the sequencing problem and the resource decision problem, and propose a multi-start iterative search heuristic (MSIS) to solve it. For the sequencing problem, an iterative search framework is constructed to effectively search the activity sequences. A two stage resource adjustment procedure and a backward peak elimination procedure is developed for solving the resource decision problem. MSIS is compared with three existing algorithms on both PSPLib and RanGen data sets involving 1380 instances. A complete calibration of the different parameters and operators of MSIS by means of a design of experiments approach is given. Experimental and statistical results show that MSIS outperforms the other three algorithms in both effectiveness and efficiency by a significant margin. (C) 2016 Published by Elsevier B.V., This work is supported by the National Natural Science Foundation of China (Nos. 61572127, 61272377), the Key Research & Development program in Jiangsu Province (No. BE2015728) and the Collaborative Innovation Center of Wireless Communications Technology. Rubén Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project SCHEYARD - Optimization of Scheduling Problems in Container Yards with reference DPI2015-65895-R co-financed with FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Performance Analysis for Heterogeneous Cloud Servers Using Queueing Theory

  • Wang, Shuang
  • Li, Xiaoping
  • Ruiz García, Rubén
[EN] In this article, we consider the problem of selecting appropriate heterogeneous servers in cloud centers for stochastically arriving requests in order to obtain an optimal tradeoff between the expected response time and power consumption. Heterogeneous servers with uncertain setup times are far more common than homogenous ones. The heterogeneity of servers and stochastic requests pose great challenges in relation to the tradeoff between the two conflicting objectives. Using the Markov decision process, the expected response time of requests is analyzed in terms of a given number of available candidate servers. For a given system availability, a binary search method is presented to determine the number of servers selected from the candidates. An iterative improvement method is proposed to determine the best servers to select for the considered objectives. After evaluating the performance of the system parameters on the performance of algorithms using the analysis of variance, the proposed algorithm and three of its variants are compared over a large number of random and real instances. The results indicate that proposed algorithm is much more effective than the other four algorithms within acceptable CPU times., This work is supported by the National Key Research and Development Program of China Grant No. 2017YFB1400801, the National Natural Science Foundation of China Grant Nos. 61572127, 61872077, 61832004 and Collaborative Innovation Center of Wireless Communications Technology. Rub~en Ruiz is partly supported by the Spanish Ministry of Science, Innovation, and Universities, under the project "OPTEP-Port Terminal Operations Optimization" (No. RTI2018-094940-BI00) financed with FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Resource Provisioning for Task-Batch Based Workflows with Deadlines in Public Clouds

  • Cai, Zhicheng
  • Li, Xiaoping
  • Ruiz García, Rubén
[EN] To meet the dynamic workload requirements in widespread task-batch based workflow applications, it is important to design algorithms for DAG-based platforms (such as Dryad, Spark and Pegasus) to rent virtual machines from public clouds dynamically. In terms of depths and functionalities, tasks of different task-batches are merged into task-units. A unit-aware deadline division method is investigated for properly dividing workflow deadlines to task deadlines so as to minimize the utilization of rented intervals. A rule-based task scheduling method is presented for allocating tasks to time slots of rented Virtual Machines (VMs) with a task right shifting operation and a weighted priority composite rule. A Unit-aware Rule-based Heuristic (URH) is proposed for elastically provisioning VMs to task-batch based workflows to minimize the rental cost in DAG-based cloud platforms. Effectiveness of the proposed URH methods is verified by comparing them against two adapted existing algorithms for similar problems on some realistic workflows., The authors would like to thank the reviewers for their constructive and useful comments. This work is supported by the National Natural Science Foundation of China (Grant No.61602243 and 61572127), the Natural Science Foundation of Jiangsu Province (Grant No.BK20160846), the Jiangsu Key Laboratory of Image and Video Understanding for Social Safety (Grant No. 30916014107). Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project "SCHEYARD" (DPI2015-65895-R) financed by FEDER funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

An Iterated Greedy Heuristic for Mixed No-Wait Flowshop Problems

  • Wang, Yamin
  • Li, Xiaoping
  • Ruiz García, Rubén
  • Sui, S.
[EN] The mixed no-wait flowshop problem with both wait and no-wait constraints has many potential real-life applications. The problem can be regarded as a generalization of the traditional permutation flowshop and the no-wait flowshop. In this paper, we study, for the first time, this scheduling setting with makespan minimization. We first propose a mathematical model and then we design a speed-up makespan calculation procedure. By introducing a varying number of destructed jobs, a modified iterated greedy algorithm is proposed for the considered problem which consists of four components: 1) initialization solution construction; 2) destruction; 3) reconstruction; and 4) local search. To further improve the intensification and efficiency of the proposal, insertion is performed on some neighbor jobs of the best position in a sequence during the initialization, solution construction, and reconstruction phases. After calibrating parameters and components, the proposal is compared with five existing algorithms for similar problems on adapted Taillard benchmark instances. Experimental results show that the proposal always obtains the best performance among the compared methods., This work was supported in part by the National Natural Science Foundation of China under Grant 61572127 and 61272377, in part by the Key Research and Development Program in Jiangsu Province under Grant BE2015728, and in part by the Collaborative Innovation Center of Wireless Communications Technology. The work of R. Ruiz was supported in part by the Spanish Ministry of Economy and Competitiveness through the project "SCHEYARD-Optimization of Scheduling Problems in Container Yards" under Grant DPI2015-65895-R, and in part by the FEDER Funds.
RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Publicación Artículo científico (JournalArticle)

Price forecasting for spot instances in Cloud computing

  • Cai, Zhicheng
  • Li, Xiaoping
  • Ruiz García, Rubén
  • Li, Qianmu
[EN] Big data applications usually need to rent a large number of virtual machines from Cloud computing providers. As a result of the policies employed by Cloud providers, the prices of spot virtual machine instances behavior stochastically. Spot prices (prices of spot instances) fluctuate greatly or have multiple regimes. Choosing virtual machines according to trends in prices is helpful in decreasing the resource rental cost. Existing price prediction methods are unable to accurately predict prices in these environments. As a result, a dynamic-ARIMA and two markov regime-switching autoregressive model based forecasting methods have been developed in this paper. Experimental results show that the proposals are better than the existing MonthAR in most scenarios. (C) 2017 Elsevier B.V. All rights reserved., The authors would like to thank the reviewers for their constructive and useful comments. This work is supported by the National Natural Science Foundation of China (Grant No. 61602243 and No. 61572127), the Natural Science Foundation of Jiangsu Province (Grant No. BK20160846), Jiangsu Key Laboratory of Image and Video Understanding for Social Safety (Grant No. 30916014107). Ruben Ruiz is partially supported by the Spanish Ministry of Economy and Competitiveness, under the project "SCHEYARD" (No. DPI2015-65895-R) financed by FEDER funds.