An Integrated Industrial Internet of Things-Optimization Framework for Data-Driven Smart Manufacturing
Industrial Internet of Things (IIoT) deployments generate high-volume operational data, yet many manufacturing systems still use this data mainly for monitoring rather than closed-loop decision support. This paper presents an integrated IIoT-optimization framework for data-driven smart manufacturing. The contribution is positioned as a conceptual and simulation-based architecture rather than a fully validated industrial deployment. The framework contains four layers: physical data acquisition, analytics and processing, optimization and decision support, and presentation/application feedback. Its Adaptive Model Integrator (AMI) is implemented as a context-aware selection policy that chooses among mathematical programming, heuristic dispatching, multi-objective optimization, model predictive control, and metaheuristic solvers according to data quality, system stability, problem structure, active objectives, and available computation time. A 720-hour synthetic MATLAB simulation of a five-line, 30-machine production system with 240 virtual sensors is used to illustrate the framework. Under the stated assumptions, the AMI-enabled configuration improves simulated throughput from 852.2 to 961.0 units/hour, reduces energy consumption from 1.198 to 1.128 kWh/unit, increases quality rate from 94.46% to 95.75%, and raises OEE from 77.39% to 84.75%. These results are reported as preliminary simulation outcomes, not as evidence of field-level industrial benefit. The paper also specifies the simulation assumptions, baseline logic, AMI decision rules, limitations, and future validation requirements.