High-level reasoning while low-level actuation in Cyber-Physical Systems: How efficient is it?

The increasing complexity of industrial information-integration systems demands software technologies that enable intelligent behaviour, real-time response, and efficient development. Although many programming languages and frameworks exist, engineers still lack sufficient empirical evidence to guide the choice of tools for advanced industrial applications. This study addresses that need by measuring and comparing worst-case execution time (WCET) and development time across six languages and frameworks: C++, Java, Jade, Jason, and fuzzy Jason BDI with both loosely and tightly coupled integration. These technologies reflect a progression from procedural and object-oriented programming to agent-based frameworks capable of symbolic and fuzzy reasoning. Rather than relying on broad concepts such as paradigms or orientations, the study adopts a developer-centred approach grounded in measurable outcomes. The structured comparison examines how rising abstraction levels and reasoning capabilities affect both development effort and runtime behaviour. By analysing these dimensions, the study highlights concrete trade-offs between engineering workload and execution efficiency. The findings show how abstraction and reasoning mechanisms shape system performance and developer productivity, offering practical insight for designing intelligent, agent-based solutions that must operate under real-time constraints and complex decision-making requirements. Overall, the study contributes evidence-based guidance for selecting software technologies in industrial informatization, supporting improved integration efficiency, maintainability, and responsiveness, and laying groundwork for future research on the interplay between language features, development dynamics, and runtime behaviour in cyber-physical and smart manufacturing systems.