{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,22]],"date-time":"2026-07-22T04:35:35Z","timestamp":1784694935988,"version":"3.55.0"},"reference-count":93,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["62572433"],"award-info":[{"award-number":["62572433"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["U25D9020"],"award-info":[{"award-number":["U25D9020"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["62201504"],"award-info":[{"award-number":["62201504"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"name":"National Key Research and Development Program of China","award":["2024YFE0200700"],"award-info":[{"award-number":["2024YFE0200700"]}]},{"DOI":"10.13039\/501100001809","name":"Intelligent Network Optimization Algorithms at the Edge","doi-asserted-by":"publisher","award":["ZFKY20250912001"],"award-info":[{"award-number":["ZFKY20250912001"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Netw."],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/ton.2026.3677371","type":"journal-article","created":{"date-parts":[[2026,3,25]],"date-time":"2026-03-25T19:54:58Z","timestamp":1774468498000},"page":"4637-4651","source":"Crossref","is-referenced-by-count":1,"title":["MoE\n                    <sup>2<\/sup>\n                    : Optimizing Collaborative Inference for Edge Large Language Models"],"prefix":"10.1109","volume":"34","author":[{"ORCID":"https:\/\/orcid.org\/0009-0003-2454-6577","authenticated-orcid":false,"given":"Lyudong","family":"Jin","sequence":"first","affiliation":[{"name":"Zhejiang University&#x2014;University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yanning","family":"Zhang","sequence":"additional","affiliation":[{"name":"Zhejiang University&#x2014;University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0003-6291-8132","authenticated-orcid":false,"given":"Yanhan","family":"Li","sequence":"additional","affiliation":[{"name":"Zhejiang University&#x2014;University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shurong","family":"Wang","sequence":"additional","affiliation":[{"name":"Zhejiang University&#x2014;University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0256-2416","authenticated-orcid":false,"given":"Howard H.","family":"Yang","sequence":"additional","affiliation":[{"name":"Zhejiang University&#x2014;University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3230-6392","authenticated-orcid":false,"given":"Jian","family":"Wu","sequence":"additional","affiliation":[{"name":"Zhejiang Key Laboratory of Medical Imaging Artificial Intelligence, Zhejiang University, Hangzhou, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4893-6946","authenticated-orcid":false,"given":"Meng","family":"Zhang","sequence":"additional","affiliation":[{"name":"Zhejiang University&#x2014;University of Illinois at Urbana-Champaign Institute, Zhejiang University, Haining, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref2","article-title":"GPT-4 technical report","volume-title":"arXiv:2303.08774","author":"Achiam","year":"2023"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2011"},{"key":"ref4","first-page":"240:1","article-title":"PaLM: Scaling language modeling with pathways","author":"Chowdhery","year":"2022","journal-title":"J. Mach. Learn. Res."},{"key":"ref5","first-page":"8469","article-title":"PaLM-E: An embodied multimodal language model","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Driess"},{"key":"ref6","article-title":"PaLM 2 technical report","volume-title":"arXiv:2305.10403","author":"Anil","year":"2023"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1991.3.1.79"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1994.6.2.181"},{"key":"ref9","article-title":"Outrageously large neural networks: The sparsely-gated mixture-of-experts layer","volume-title":"arXiv:1701.06538","author":"Shazeer"},{"key":"ref10","article-title":"GShard: Scaling giant models with conditional computation and automatic sharding","volume-title":"arXiv:2006.16668","author":"Lepikhin"},{"issue":"120","key":"ref11","first-page":"1","article-title":"Switch transformers: Scaling to trillion parameter models with simple and efficient sparsity","volume":"23","author":"Fedus","year":"2022","journal-title":"J. Mach. Learn. Res."},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2019.8737560"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/TVT.2018.2881191"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/BID.2017.8336572"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2017.2745201"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/TWC.2019.2946140"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2922285"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2017.2682318"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2024.3415661"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1016\/j.comnet.2022.109150"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM48880.2022.9796896"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1007\/s12083-021-01223-1"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2022.3218724"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2018.1700202"},{"key":"ref25","article-title":"LocMoE: A low-overhead MoE for large language model training","author":"Li","year":"2024","journal-title":"arXiv:2401.13920"},{"key":"ref26","article-title":"EdgeMoE: Empowering sparse large language models on mobile devices","author":"Yi","year":"2023","journal-title":"arXiv:2308.14352"},{"key":"ref27","article-title":"JetMoE: Reaching Llama2 performance with 0.1M dollars","author":"Shen","year":"2024","journal-title":"arXiv:2404.07413"},{"key":"ref28","article-title":"Faster, smaller, and smarter: Task-aware expert merging for online MoE inference","author":"Han","year":"2025","journal-title":"arXiv:2509.19781"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2025.3527641"},{"key":"ref30","article-title":"Not all patches are what you need: Expediting vision transformers via token reorganizations","author":"Liang","year":"2022","journal-title":"arXiv:2202.07800"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.825"},{"key":"ref32","article-title":"Cached model-as-a-resource: Provisioning large language model agents for edge intelligence in space-air-ground integrated networks","author":"Xu","year":"2024","journal-title":"arXiv:2403.05826"},{"key":"ref33","article-title":"Hybrid LLM: Cost-efficient and quality-aware query routing","volume-title":"arXiv:2404.14618","author":"Ding"},{"key":"ref34","article-title":"ServerlessLLM: Low-latency serverless inference for large language models","author":"Fu","year":"2024","journal-title":"arXiv:2401.14351"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/VTC2023-Fall60731.2023.10333824"},{"key":"ref36","first-page":"3690","article-title":"Power-BERT: Accelerating BERT inference via progressive word-vector elimination","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Goyal"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/JSAC.2021.3126087"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/ICCC57788.2023.10233481"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/TWC.2022.3221778"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1269"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1145\/3603269.3610856"},{"key":"ref42","first-page":"19274","article-title":"Fast inference from transformers via speculative decoding","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Leviathan"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1145\/3552326.3587438"},{"key":"ref44","article-title":"Large language model (LLM) for telecommunications: A comprehensive survey on principles, key techniques, and opportunities","author":"Zhou","year":"2024","journal-title":"arXiv:2405.10825"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1145\/3639856.3639892"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2837"},{"key":"ref47","article-title":"The power of large language models for wireless communication system development: A case study on FPGA platforms","author":"Du","year":"2023","journal-title":"arXiv:2307.07319"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1145\/3626111.3628183"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1145\/3626111.3628189"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.23919\/DATE56975.2023.10137086"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.23919\/CNSM59352.2023.10327837"},{"key":"ref52","article-title":"Making network configuration human friendly","author":"Wang","year":"2023","journal-title":"arXiv:2309.06342"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1145\/3626111.3628194"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-25538-0_3"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.3390\/jcp1040031"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2020.106529"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2024.3363469"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2022.3185748"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1109\/GLOBECOM54140.2023.10437725"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.3390\/brainsci13010147"},{"key":"ref61","article-title":"Visual classification via description from large language models","author":"Menon","year":"2022","journal-title":"arXiv:2210.07183"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01438"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.3390\/electronics12030516"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1145\/3485447.3512217"},{"key":"ref65","article-title":"Self-refined large language model as automated reward function designer for deep reinforcement learning in robotics","author":"Song","year":"2023","journal-title":"arXiv:2309.06687"},{"key":"ref66","article-title":"Reward design with language models","author":"Kwon","year":"2023","journal-title":"arXiv:2303.00001"},{"key":"ref67","article-title":"Eureka: Human-level reward design via coding large language models","volume-title":"arXiv:2310.12931","author":"Ma"},{"key":"ref68","article-title":"Towards optimizing with large language models","author":"Guo","year":"2023","journal-title":"arXiv:2310.05204"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1080\/03155986.2024.2385189"},{"key":"ref70","article-title":"OptiMUS: Optimization modeling using MIP solvers and large language models","author":"AhmadiTeshnizi","year":"2023","journal-title":"arXiv:2310.06116"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1145\/3583133.3596401"},{"key":"ref72","article-title":"Large language model for multi-objective evolutionary optimization","author":"Liu","year":"2023","journal-title":"arXiv:2310.12541"},{"key":"ref73","article-title":"TimeGPT-1","author":"Garza","year":"2023","journal-title":"arXiv:2310.03589"},{"key":"ref74","article-title":"A decoder-only foundation model for time-series forecasting","volume-title":"arXiv:2310.10688","author":"Das"},{"key":"ref75","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2023.3342137"},{"key":"ref76","article-title":"Time-LLM: Time series forecasting by reprogramming large language models","volume-title":"arXiv:2310.01728","author":"Jin"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1145\/3719207"},{"key":"ref78","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1877"},{"key":"ref79","article-title":"Meta-transformer: A unified framework for multimodal learning","author":"Zhang","year":"2023","journal-title":"arXiv:2307.10802"},{"key":"ref80","doi-asserted-by":"publisher","DOI":"10.1145\/3651890.3672268"},{"key":"ref81","article-title":"Theory of mixture-of-experts for mobile edge computing","author":"Li","year":"2024","journal-title":"arXiv:2412.15690"},{"key":"ref82","first-page":"31094","article-title":"FlexGen: High-throughput generative inference of large language models with a single GPU","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Sheng"},{"key":"ref83","article-title":"MoE2: Optimizing collaborative inference for edge large language models","author":"Jin","year":"2025","journal-title":"arXiv:2501.09410"},{"key":"ref84","doi-asserted-by":"publisher","DOI":"10.1126\/science.127.3293.292.a"},{"key":"ref85","doi-asserted-by":"publisher","DOI":"10.1016\/0893-6080(89)90020-8"},{"key":"ref86","article-title":"Discrete monotonic global optimization","author":"Minoux","year":"2002"},{"key":"ref87","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-industry.103"},{"key":"ref88","article-title":"Qwen2.5 technical report","volume-title":"arXiv:2412.15115","author":"Yang","year":"2024"},{"key":"ref89","article-title":"The llama 3 herd of models","author":"Grattafiori","year":"2024","journal-title":"arXiv:2407.21783"},{"key":"ref90","article-title":"Mistral 7B","author":"Jiang","year":"2023","journal-title":"arXiv:2310.06825"},{"key":"ref91","article-title":"Measuring massive multitask language understanding","author":"Hendrycks","year":"2020","journal-title":"arXiv:2009.03300"},{"key":"ref92","article-title":"More agents is all you need","author":"Li","year":"2024","journal-title":"arXiv:2402.05120"},{"key":"ref93","volume-title":"VLLM: Easy, Fast, and Cheap LLM Serving With Pagedattention","author":"Kwon et al","year":"2023"}],"container-title":["IEEE Transactions on Networking"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10723154\/11317935\/11456235.pdf?arnumber=11456235","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T19:52:21Z","timestamp":1776455541000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11456235\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":93,"URL":"https:\/\/doi.org\/10.1109\/ton.2026.3677371","relation":{},"ISSN":["2998-4157"],"issn-type":[{"value":"2998-4157","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}