Despite the continuous development of advanced control algorithms, PID-based controllers, including two-degree-of-freedom proportional-integral (2-DOF PI) structures, dominate in industrial practice. However, obtaining optimal tuning parameters remains a significant application challenge. Contemporary metaheuristic approaches typically offload computations to higher-level supervisory layers, which introduces non-deterministic communication delays and cybersecurity vulnerabilities. Addressing this issue, this paper presents a fully autonomous, Differential Evolution (DE)-based autotuning strategy executed natively on a Programmable Logic Controller (PLC) at the field level. By explicitly accounting for strict memory and cycletime constraints, a universal, hardware-optimized function block was developed. This libraryconformant solution preserves deterministic controller operation and compatibility with standard industrial workflows. Resource analysis on standard PLCs confirmed the implementation feasibility. Furthermore, industrial validation through the virtual commissioning of a heat generation and distribution system demonstrated that the proposed algorithm outperforms commercial autotuning functions and the widely adopted Skogestad Internal Model Control (SIMC) strategy. This work proves that metaheuristics can be safely, effectively, and natively deployed in the field-level process control layer, eliminating the need for external computing units.