Markov Task Network: A Framework for Service Composition under Uncertainty in Cyber-Physical Systems.

Area of Science:

• Computer Science
• Artificial Intelligence
• Systems Engineering

Background:

• Collaborative systems face challenges integrating diverse cyber and physical components.
• Dynamic formation of collaborative services is crucial for achieving system goals in pervasive cyber-physical systems.

Purpose of the Study:

• To propose a cross-layer automation and management model for cyber-physical systems.
• To address the dynamic formation of collaborative services and manage uncertainty.

Main Methods:

• Developed an ontology-oriented hierarchical task network model for dynamic service formation.
• Introduced a novel Markov task network, bridging hierarchical task networks and Markov logic networks.
• Utilized semantic reasoning for dynamic task composition from high-level goals.

Main Results:

• The proposed model enables dynamic composition of primitive tasks from high-level system goals.
• The Markov task network reduces computational and inferential loads in task decomposition.
• Achieved high-precision service composition even under uncertain conditions.

Conclusions:

• The ontology-oriented hierarchical task network model effectively supports dynamic service formation in cyber-physical systems.
• The Markov task network offers an efficient approach to handle uncertainty in task decomposition.
• This research provides a robust framework for advanced collaborative cyber-physical systems.