Swarm and Evolutionary Computation最新文献

{"title":"Learning to sample in variable neighborhood search algorithm for urban cable routing optimization","authors":"Wei Liu ,&nbsp;Rui Wang ,&nbsp;Chenhui Lin ,&nbsp;Kaiwen Li ,&nbsp;Wenhua Li ,&nbsp;Tao Zhang","doi":"10.1016/j.swevo.2026.102432","DOIUrl":"10.1016/j.swevo.2026.102432","url":null,"abstract":"<div><div>Urban underground cable construction is essential for enhancing power grid reliability, yet the high construction costs demand systematic optimization. Constrained by road network infrastructure, this optimization problem requires consideration not only of connectivity relationships between substations but also of specific routing strategies along road networks, constituting a large-scale bilevel combinatorial optimization problem. Insufficient attention to specific routing subproblems in traditional research and simplistic algorithmic designs that are ill-equipped for large-scale combinatorial optimization leave substantial room for advancement in addressing this complex optimization challenge. To navigate the enormous combinatorial search space, we propose a learning-assisted variable neighborhood search (L-VNS) algorithm integrating four key components. First, an auxiliary task focusing exclusively on the upper-level connectivity subproblem generates high-quality initial solutions by employing hybrid genetic search for connection optimization and A* for detailed path routing. Subsequently, the algorithm iteratively refines the connectivity topology using variable neighborhood search equipped with three complementary operators. A multi-agent deep reinforcement learning module adaptively guides probabilistic neighborhood sampling by jointly encoding both upper-level connectivity patterns and lower-level routing structures, effectively exploiting problem structure. Finally, a modified A* operator re-plans the lower-level paths affected by neighborhood modifications to ensure feasibility and solution completeness. Comprehensive experiments on 12 benchmark instances and 3 GIS-derived instances demonstrate the superiority of L-VNS, achieving total construction cost reductions of 0.92–73.72% compared to representative approaches. Ablation studies and sensitivity analyses further validate the effectiveness and robustness of the proposed algorithm.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102432"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179796","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Dynamic Probability-Controlled Learning Swarm Optimizer for large-scale optimization","authors":"Nengxian Liu ,&nbsp;Yingqiang Li ,&nbsp;Mingjian Fu ,&nbsp;Genggeng Liu ,&nbsp;Jeng-Shyang Pan","doi":"10.1016/j.swevo.2026.102424","DOIUrl":"10.1016/j.swevo.2026.102424","url":null,"abstract":"<div><div>Large-scale optimization problems (LSOPs) pose significant challenges, as their vast search spaces often lead to a trade-off dilemma between effective exploration and efficient exploitation. To address this issue, this paper proposes a Dynamic Probability-Controlled Learning Swarm Optimizer (DPCLSO). DPCLSO introduces two coordinated mechanisms: (1) a probability-controlled learning mechanism, which adaptively regulates particle updates based on individual quality and search stage; and (2) a dynamic learning sample construction strategy, which provides more suitable learning directions for particles with different fitness levels. By synergizing these mechanisms, DPCLSO aims to achieve a more effective exploration–exploitation balance throughout the search process. Extensive experiments on the CEC2010 and CEC2013 benchmark suites demonstrate the superiority of DPCLSO, achieving the highest overall ranking in the Friedman test on both suites. In particular, it obtains superior or competitive results on at least 14 out of 20 functions in the CEC2010 benchmark. These results demonstrate that DPCLSO is a competitive and effective alternative for large-scale optimization problems.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102424"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179847","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Implementing automated task scheduling framework for fog network using adaptive actor critic deep reinforcement learning with sparse attention through multi-objective constraints","authors":"Rayroth Shashank ,&nbsp;Aswin Kumer S V ,&nbsp;Selvaganapathi Sennan ,&nbsp;Thankaraj A ,&nbsp;Britto Raj S ,&nbsp;E. Mohan","doi":"10.1016/j.swevo.2026.102417","DOIUrl":"10.1016/j.swevo.2026.102417","url":null,"abstract":"<div><div>The rapid advancement of the Internet of Things (IoT) has led to massive data generation, requiring efficient processing strategies. Immediate-response tasks are handled at fog nodes due to their proximity, while complex computations are offloaded to cloud data centers, offering extensive storage and processing power. Fog computing has emerged as a promising approach for real-time execution, enabling low-latency services at the network edge. However, constrained resources in fog environments make task scheduling challenging. While fog computing reduces transmission delays, it increases energy consumption, whereas cloud offloading minimizes energy usage but introduces higher latency. To address these challenges, this work proposes an advanced scheduling framework integrating Adaptive Actor-Critic Deep Reinforcement Learning with Sparse Attention (A-ACDRL-SA) for optimal task allocation. Intellectual Randomized Conceptualization-based Dollmaker Optimization Algorithm (IRC-DOA) fine-tunes scheduling parameters, enhancing efficiency. Multi-objective constraints such as response time, CPU load, processing delay, latency, and storage utilization are optimized. The framework dynamically adapts to workload variations, ensuring efficient resource management and reliable execution. Experimental evaluations demonstrate superior performance compared to classical methods, achieving reduced latency, optimized energy consumption, and scalability. This intelligent task scheduling technique enhances network reliability while offering sustainable solutions for cloud-fog computing environments.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102417"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dual-path information bottleneck-enhanced heterogeneous graph attention networks for dynamic flexible job-shop scheduling problem","authors":"Fangyu Li,&nbsp;Jiarong Du,&nbsp;Jingjing Wang,&nbsp;Honggui Han","doi":"10.1016/j.swevo.2026.102431","DOIUrl":"10.1016/j.swevo.2026.102431","url":null,"abstract":"<div><div>Dynamic flexible job shop scheduling problem (DFJSP) plays a crucial role in enhancing the productivity of flexible manufacturing systems. Recent works have employed deep reinforcement learning to solve DFJSP due to the complexity of the scheduling environment. However, existing studies neglect the importance of scheduling stability, which is essential for generalization in dynamic environments. We propose a dual-path IB-enhanced heterogeneous graph attention networks (DIG) to solve DFJSP, which constructs a heterogeneous graph architecture to embed state features and introduces an information bottleneck (IB) to compress redundant information in the embedding. First, we design a heterogeneous graph attention network (HGAT) to embed the heterogeneous disjunction graph nodes, facilitating adaptive scheduling in dynamic environments. Second, we exert IB regularization to HGAT, optimizing the balance between state space embedding and compressed representations to achieve stability in dynamic environments. Third, we build a graph-based end-to-end dynamic reinforcement learning framework, which ensures the stability of the scheduling method and improves the scheduling efficiency. The experiment results indicate that DIG achieves significantly shorter processing times and exhibits greater stability when compared with both classical scheduling algorithms and the SOTA scheduling methods.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102431"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148242414","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing the scalability of large-scale multi-objective evolutionary algorithm through frequency domain search","authors":"Wenxiang Wang ,&nbsp;Zhiping Tan ,&nbsp;Yi Wang ,&nbsp;Weiping Zhang","doi":"10.1016/j.swevo.2026.102423","DOIUrl":"10.1016/j.swevo.2026.102423","url":null,"abstract":"<div><div>In the optimization of large-scale problems, the scalability of the algorithm is very important. In the frequency domain analysis of physical signals, the frequency domain dimension and the time domain dimension are independent of each other. Inspired by this, this paper designs a frequency domain search-based evolutionary algorithm (FDSEA) for large-scale multi-objective optimization problems. The algorithm adopts a dual-population cooperative framework: the frequency domain subpopulation extracts the global features of the optimization problem by searching the low-frequency parameters of the decision variables in the frequency domain, to accelerate the convergence of the algorithm; while the decision domain subpopulation supplements the search in the decision domain to improve the convergence accuracy of the algorithm, and uses the adaptive evolution operator based on dual-activation functions to realize rapid and dynamic operator switching. Two subpopulations share evolution information in real time to realize coevolution. Due to the complete decoupling and independence between the dimensions of the frequency domain and decision domain, the algorithm has good scalability. Numerical experiments show that FDSEA not only has better comprehensive performance than the five state-of-the-art algorithms, but also shows excellent scalability in optimization experiments within 10,000 dimensions.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102423"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179803","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Advanced Self-Adaptive Distributed Differential Evolution with Hierarchical Learning algorithm for multimodal optimization in communication system","authors":"Arup Kumar Ghosh ,&nbsp;Subhankar Bhattacharjee ,&nbsp;Gautam Garai","doi":"10.1016/j.swevo.2026.102420","DOIUrl":"10.1016/j.swevo.2026.102420","url":null,"abstract":"<div><div>Modern communication systems require advanced optimization algorithms capable of navigating high-dimensional, multimodal design spaces to achieve spectral precision, energy efficiency, and robust fabrication tolerance. This paper introduces <em>Advanced Self-Adaptive Distributed Differential Evolution with Hierarchical Learning</em> (ASADDE-HL), a novel algorithm tailored for multimodal optimization of photonic components, including dense wavelength division Multiplexers/Demultiplexers (MUX/DEMUX). The algorithm combines three key innovations: (1) Hierarchical Subspace Decomposition and Cooperative Optimization, (2) Multi-Granularity Self-Adaptive Learning, and (3) Integrated Three-Phase Search Strategy. These features collectively promote a balanced exploration–exploitation trade-off and empirically improve convergence speed and robustness in dynamic optimization landscapes. Rigorous evaluations on the CEC 2017 benchmark functions and real-world photonic design problems demonstrate the superiority of ASADDE-HL over state-of-the-art algorithms such as self-adaptive niching differential evolution (SANDE), adaptive distributed differential evolution with hierarchy learning (ADDE-HL), multi-strategies differential evolution with Q-learning (MSDE-QL), and hybrid cultural algorithm-differential evolution (HCADE). The algorithm exhibits exceptional performance, achieving a transmission efficiency of 99.9648% and a Full Width at Half Maximum (FWHM) of only 0.4 nm in the optimized filters. Moreover, no crosstalk is observed between adjacent transmission windows in the multiplexed output, and the design maintains less than 6% performance degradation under <span><math><mrow><mo>±</mo><mn>7</mn><mtext>%</mtext></mrow></math></span> fabrication variations. Its adaptive learning mechanisms and scalable distributed architecture make it ideal for real-world photonic optimization. ASADDE-HL demonstrates competitive and superior performance compared to several state-of-the-art evolutionary algorithms in photonic device optimization for advanced communication systems.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102420"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179848","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diffusion evolutionary algorithm for neural architecture search","authors":"Yu Xue ,&nbsp;Pingyuan Cai ,&nbsp;Ferrante Neri ,&nbsp;Yong Zhang ,&nbsp;Dunwei Gong","doi":"10.1016/j.swevo.2026.102421","DOIUrl":"10.1016/j.swevo.2026.102421","url":null,"abstract":"<div><div>Traditional Neural Architecture Search (NAS) methods often struggle with hyperparameter tuning and are prone to local optima, which limits the diversity of the population and hampers the search for optimal architectures. This paper introduces DiffEvo-NAS, a novel approach that leverages a diffusion evolution algorithm, which interprets the evolution process as a denoising operation and the reverse evolution as diffusion. By incorporating directional denoising (similar to directional selection) and adding small noise (analogous to mutations), the algorithm continuously refines individuals while maintaining diversity, enabling effective exploration of the search space and avoiding local optima. The effectiveness of DiffEvo-NAS is demonstrated through experiments on the CIFAR-10 and CIFAR-100 datasets, where it achieves a test error rate of 2.55% and 16.15%, respectively. On ImageNet, it achieves competitive performance with a top-1 error rate of 24.61% and top-5 error rate of 7.44%. These results confirm that DiffEvo-NAS offers significant improvements in both exploration and exploitation, outperforming other evolutionary NAS methods in terms of population diversity and architecture performance. As a stepping stone in the field, DiffEvo-NAS lays the groundwork for further innovations in NAS, particularly in the integration of diffusion models and evolutionary computation.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102421"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179849","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Research on UAV-USV cooperative pursuit method based on deep reinforcement learning","authors":"Shoucong Wang ,&nbsp;Yong Xie ,&nbsp;Xiaobo Liu","doi":"10.1016/j.swevo.2026.102433","DOIUrl":"10.1016/j.swevo.2026.102433","url":null,"abstract":"<div><div>UAV-USV cooperative pursuit is crucial for maintaining maritime security. The Multi-Agent Deep Deterministic Policy Gradient (MADDPG) is a classical algorithm for solving multi-agent cooperative pursuit problems. However, due to differences in observation capabilities and motion dimensions between UAVs and USVs, the input distribution becomes inconsistent, which increases the learning difficulty of the centralized Critic network, thereby reducing the scalability and convergence speed of MADDPG. Based on MADDPG, this paper proposes a Distributed Multi-Agent Deep Deterministic Policy Gradient algorithm incorporating Sum-Tree-based Prioritized Experience Replay (STPER-DMADDPG). To improve scalability, the proposed method decouples the centralized Critic network, thereby reducing its reliance on joint state–action information while incorporating an information-sharing mechanism to enhance coordination among heterogeneous agents. Meanwhile, we employ STPER to prioritize important transitions, which accelerates network convergence and improves training stability. Simulation results show that the proposed method outperforms baseline algorithms in convergence speed, average reward, and task completion efficiency. Furthermore, it exhibits stability and scalability as the number of agents increases.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102433"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148242413","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Particle Swarm Optimization with complementary velocity learning scheme and multiple performance mechanisms: Applications in power electronics","authors":"Libin Hong ,&nbsp;Chengbo Dai ,&nbsp;Tianxiang Cui ,&nbsp;Zheng Lu ,&nbsp;Ning Xue","doi":"10.1016/j.swevo.2026.102410","DOIUrl":"10.1016/j.swevo.2026.102410","url":null,"abstract":"<div><div>The integration of multiple operators and mechanisms in metaheuristics is an effective methodology for improving their performance. Meanwhile, to further enhance the level of intelligence in metaheuristics, both the complexity of the target problem and learning strategies should be considered when designing metaheuristic algorithms. In this work, a Complementary velocity learning scheme, an Improved estimation of distribution algorithm, Information archiving, and the Interior-point method based Particle Swarm Optimization (CIII-PSO) is proposed. The complementary velocity learning scheme not only combines three different velocity update strategies with distinct exploration and exploitation capabilities, but also integrates learning strategies to better guide particles during swarm-based search. The improved estimation of distribution algorithm accounts for both the complexity of the target problem and the enhancement of exploitation. The information archiving mechanism generates both parameters and archived particles, providing rich options and historical information (parameters and particles from early iterations) for learning. The interior-point method is a local search mechanism that explores the neighborhood of the best particles, thereby greatly enhancing the exploitation capability of CIII-PSO. CIII-PSO is evaluated against 15 recently proposed PSO-based algorithms on the IEEE Congress on Evolutionary Computation 2013 and 2017 (CEC2013 and CEC2017) benchmark suites, demonstrating superior performance across all test functions. Ablation studies further verify the effectiveness and individual contributions of the proposed components. In addition, CIII-PSO is applied to power electronics problems, achieving outstanding performance in the experiments. The source code of CIII-PSO is provided at <span><span>https://github.com/microhard1999/CODES</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102410"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Hesitant fuzzy multi-attribute decision making with swarm optimization for sustainable supply chain network design","authors":"Muhammad Tahir ,&nbsp;Wajahat Ali ,&nbsp;Aneesa Nawaz","doi":"10.1016/j.swevo.2026.102418","DOIUrl":"10.1016/j.swevo.2026.102418","url":null,"abstract":"<div><div>For a sustainable system, in this paper, two new aggregation operators in the hesitant fuzzy (HF) system are suggested: the TA exponential root operator (TA-ERO) and the TA logarithmic-sigmoid operator (TA-LSO), which are aimed at overcoming key shortcomings of the current alternatives, such as the low discrimination power and lack of non-linear relationship modeling. These operators are exponential functions combined with logarithmic transformations and sigmoid activations as well as nested root structures, which are used to capture fine variation and saturation of supply chain networks. Strict mathematical demonstrations prove fundamental properties including closure, idempotency (exact for TA-ERO, approximate for TA-LSO), monotonicity, boundedness, and commutativity. An all-encompassing multi-attribute decision-making (MADM) system is created with multi-expert evaluation, distance-based consistency checking, and adaptable operator choice. The usefulness and excellence of the suggested methodology are proved using a case study of swarm optimization of a sustainable supply chain network with 9 alternatives, 6 criteria, and 3 experts. The findings indicate that TA-ERO has a better score range of 0.0835 (calculated as <span><math><mrow><mn>0</mn><mo>.</mo><mn>9751</mn><mtext>–</mtext><mn>0</mn><mo>.</mo><mn>8916</mn></mrow></math></span> from the case study results), as opposed to 0.081 for current HFWA and HFWG operators, representing approximately 3% improved discrimination performance. Sensitivity analysis establishes the exceptional robustness of seven alternatives exhibiting perfect rank invariance (<span><math><mrow><mi>Δ</mi><msub><mrow><mi>R</mi></mrow><mrow><mi>i</mi></mrow></msub><mo>=</mo><mn>0</mn></mrow></math></span>) and the best alternative with rank 1 in all five weight situations. The comparative analysis provides evidence of the perfect ranking consistency and helps to prove the better discriminatory power of TA-ERO. The results confirm that the suggested TA operators allow offering mathematically valid, practically tested, and demonstrably better tools for HF decision-making in complex supply chain contexts.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"106 ","pages":"Article 102418"},"PeriodicalIF":8.5,"publicationDate":"2026-06-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148179790","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}