This paper comprehensively surveys the intersection of distributed intelligence and model optimization in Edge-Cloud Collaborative Computing (ECCC). ECCC is a core paradigm that integrates cloud resources and edge devices to meet the computing demands of modern intelligent applications. It provides a systematic tutorial on the underlying architecture, enabling technologies, and emerging applications. It analyzes model optimization approaches such as model compression, adaptation, and neural architecture search, as well as AI-based resource management strategies that balance performance, energy efficiency, and latency requirements. Furthermore, it explores critical aspects of privacy and security enhancement within ECCC systems and examines real-world deployments across diverse applications in autonomous driving, healthcare, and industrial automation. Through performance analysis and benchmarking techniques, it establishes evaluation standards for these complex systems. It also suggests important research directions, including LLM deployment, 6G integration, neuromorphic computing, and quantum computing, and provides a roadmap for addressing heterogeneity management, real-time processing, and scalability.
