With the rapid development of EC, multimodal large models have shown important value in intelligent perception and decision-making. However, the existing multimodal model deployment schemes are difficult to meet the actual needs of edge device resources being limited and delay requirements being strict, resulting in low model reasoning efficiency, high energy consumption, and untimely response. To this end, this paper combines artificial intelligence model optimization technology with EC architecture and proposes a set of compression, quantization and distributed reasoning collaborative optimization strategies for multimodal large models, aiming to improve the operating efficiency and performance stability of the model on edge nodes. By designing multimodal data synchronous collection, feature fusion and dynamic scheduling mechanisms, the system optimizes the reasoning process and further reduces the computing burden and energy consumption. The experimental results show that the performance of the distributed deployment model is the most balanced, with a Top-1 accuracy of 91.8% and a Top-5 accuracy of 98.3%, with almost no obvious performance degradation. This shows that the distributed model sharding and node collaboration can basically retain the discrimination ability of the original model while improving efficiency, and has excellent performance stability.

Hua Yuan, Hui Chen, Ying Zhong. Deployment Optimization and Performance Testing of Multimodal Large Models in Edge Computing Environments. Academic Journal of Computing & Information Science (2025), Vol. 8, Issue 7: 121-129. https://doi.org/10.25236/AJCIS.2025.080715.

[1] Hua H, Li Y, Wang T, et al. Edge computing with artificial intelligence: A machine learning perspective[J]. ACM Computing Surveys, 2023, 55(9): 1-35.

[2] Modupe O T, Otitoola A A, Oladapo O J, et al. Reviewing the transformational impact of edge computing on real-time data processing and analytics[J]. Computer Science & IT Research Journal, 2024, 5(3): 603-702.

[3] Lu S, Lu J, An K, et al. Edge computing on IoT for machine signal processing and fault diagnosis: A review[J]. IEEE Internet of Things Journal, 2023, 10(13): 11093-11116.

[4] Sadatdiynov K, Cui L, Zhang L, et al. A review of optimization methods for computation offloading in edge computing networks[J]. Digital Communications and Networks, 2023, 9(2): 450-461.

[5] Fan W, Su Y, Liu J, et al. Joint task offloading and resource allocation for vehicular edge computing based on V2I and V2V modes[J]. IEEE Transactions on Intelligent Transportation Systems, 2023, 24(4): 4277-4292.

[6] Awad A I, Fouda M M, Khashaba M M, et al. Utilization of mobile edge computing on the Internet of Medical Things: A survey[J]. ICT Express, 2023, 9(3): 473-485.

[7] Sharif Z, Jung L T, Ayaz M, et al. Priority-based task scheduling and resource allocation in edge computing for health monitoring system[J]. Journal of King Saud University-Computer and Information Sciences, 2023, 35(2): 544-559.

[8] Mohy-Eddine M, Guezzaz A, Benkirane S, et al. An effective intrusion detection approach based on ensemble learning for IIoT edge computing[J]. Journal of Computer Virology and Hacking Techniques, 2023, 19(4): 469-481.

[9] Zaman S K, Jehangiri A I, Maqsood T, et al. LiMPO: Lightweight mobility prediction and offloading framework using machine learning for mobile edge computing[J]. Cluster Computing, 2023, 26(1): 99-117.

[10] Aghazadeh R, Shahidinejad A, Ghobaei‐Arani M. Proactive content caching in edge computing environment: A review[J]. Software: Practice and Experience, 2023, 53(3): 811-855.

[11] Yasir M, uz Zaman S K, Maqsood T, et al. CoPUP: Content popularity and user preferences aware content caching framework in mobile edge computing[J]. Cluster Computing, 2023, 26(1): 267-281.