1. Formal Approaches to Swarm Technologies Technical Briefing Christopher Rouff, Amy Vanderbilt - SAIC Walt Truszkowski, James Rash - NASA GSFC, Code 588 Mike Hinchey - NASA GSFC, Code 581 GSFC SAS 2004

2. 1 Problem Future space missions will require cooperation between multiple satellites/rovers Developers are proposing intelligent, autonomous swarms to do new science Swarm-based systems are highly parallel and nondeterministic Testing these systems using current techniques will be difficult to impossible

3. 2SAS 2004 ANTS Mission

4. 3SAS 2004 ANTS MISSION

5. 4SAS 2004 Difficulty of Testing Swarms Emergent properties that may not be known Highly distributed and parallel Large number of interacting entities Worse than exponential growth in interactions Intelligent entities (capabilities increase over time) Total or near total autonomy Very little experience in verification and validation of swarm-based systems

6. 5SAS 2004 Relevance to NASA Many future missions will be intelligent collaborative distributed systems Autonomy and autonomic properties will be necessary to manage and operate these systems Swarms will allow new way for doing science Verifying these missions will be paramount due to complexity and emergent behaviors Need verification techniques for these systems

7. 6SAS 2004 Approach Survey formal approaches for agent-based, multi- agent and swarm-based systems for appropriate swarm-based methods Apply most promising approaches to parts of ANTS Evaluate methods for needed properties Model and outline swarm-based formal method Develop formal method for swarm-based systems Do formal specification of ANTS using new method Prototype support tools

8. 7SAS 2004 Accomplishments Surveyed formal approaches Applied most promising approaches to ANTS Evaluated methods Developed properties for swarm formal method Developed model and outline for a hybrid swarm-based formal method

9. 8SAS 2004 Formal Methods Survey Criteria for Surveying Formal Methods  Emphasis given to formal methods used for: Swarm, multi-agent, or distributed specifications  Evaluated based on: Concurrency support, algorithm support, tool support, formal basis, used in agent-based specifications, used in swarm-based specifications

10. 9SAS 2004 Formal Methods Surveyed Process Algebras  CSP, CCS, Pi-Calculus, I/O Automata Model-Oriented Approaches  Z, B, FSMs, Statecharts, Petri Nets, X-Machines Logics  Temporal, Real Time, BDI, KARO Other Approaches  Artificial Physics, Mathematical Analysis, Game Theory, UML, Hybrid Approaches

11. 10SAS 2004 Specifications CSP – Communicating Sequential Processes WSCCS – Weighted Synchronous Calculus of Communicating Systems Unity logic X-Machines

12. 11SAS 2004 CSP Communicating Sequential Processes  Developed by C.A.R. Hoare  Process Algebra  Processes are recursively defined as an event followed by a process  Use channels for processes to communicate  Channels are guarded by events  Properties of CSP specifications can be proven correct

13. 12SAS 2004 CSP

14. 13SAS 2004 CSP

15. 14SAS 2004 CSP

16. 15SAS 2004 CSP

17. 16SAS 2004 CSP Evaluation  Strengths Specifying inter-process protocols Identifying race conditions Easily translated into model checking language  Weaknesses No mechanism for analyzing emergent behavior

18. 17SAS 2004 WSCCS Weighted Synchronous Calculus of Communicating Systems  Developed by Chris Tofts  Used to model social insects  Uses a state-based approach with frequencies and priorities for transitions  Frequencies and priorities can give you the probability of taking an action

19. 18SAS 2004 WSCCS Agent StateActions leading to the agent statefp Identity Communicating SendMessageWorker502 SendMessageLeader502 SendMessageError11 ReceiveMessageWorker502 ReceiveMessageLeader502 ReceiveMessageError11 Reasoning ReasoningDeliberatve502 ReasoningReactive502 Processing ProcessingSortingAndStorage172 ProcessingGeneration172 ProcessingPrediction172 ProcessingDiagnosis162 ProcessingRecovery162 ProcessingRemediation172

20. 19SAS 2004 WSCCS

21. 20SAS 2004 WSCCS Communicating Reasoning Processing

22. 21SAS 2004 WSCCS Evaluation  Strengths Process algebra strengths Specifies priorities and probabilities of actions Rules for predicting behavior  Weaknesses Ability to track and model goals and other aspects of the spacecraft

23. 22SAS 2004 Unity Logic Developed by Chandy and Misra Syntax similar to Propositional Logic Reason about predicates and states Very good for concurrent systems

24. 23SAS 2004 Unity Logic Assertions [Communicating]ReasoningDeliberatve(Leader)[Reasoning] [Reasoning] SendMessage (Leader,Worker)[Communicating] [Processing] SendMessage (Leader, Worker) [Communicating]

25. 24SAS 2004 Unity Logic Evaluation  Strengths Strengths of logic-based systems A method for reasoning about predicates and the states they imply A method for defining specific mathematical and statistical calculations to be performed  Weaknesses Not rich enough for reasoning about emergent behavior

26. 25SAS 2004 X-Machines Developed by Samuel Eilenberg Based on Finite State Machines Have an internal memory state Transitions are functions:  Input x Memory → Output x Memory

27. 26SAS 2004 X-Machines

28. 27SAS 2004 X-Machines

29. 28SAS 2004 X-Machines Set of states Transition functions

30. 29SAS 2004 X-Machines StartSendMessage Commun. ReceiveMessageCommun. ReasonReasoning ProcessProcessing CommunicatingSendMessage Commun. ReceiveMessageCommun. ReasonReasoning ProcessProcessing ReasoningSendMessage Commun. ReceiveMessageCommun. ReasonReasoning ProcessProcessing ProcessingSendMessage Commun. ReceiveMessageCommun. ReasonReasoning ProcessProcessing Leader States and Transitions

31. 30SAS 2004 X-Machines

32. 31SAS 2004 X-Machines Evaluation  Strengths State-based system with memory Executable  Weaknesses No robust means for reasoning about or predicting behaviors beyond standard propositional logic

33. 32SAS 2004 Model of Formal Method Blending of methods seems to be the best approach for specifying swarm-based systems Blending of priority and probability aspects of WSCCS with memory and transition function aspects of X-Machines as well as logic from Unity Logic is a possibility and is currently being researched

34. 33SAS 2004 Model of Formal Method

35. 34SAS 2004 Future Work Exploring Chaos Theory Modifying WSCCS with different probabilistic methods Blending above methods into a hybrid formal method

36. 35SAS 2004 Next steps Develop formal method for swarm-based systems based on current model Do formal specification of ANTS using new method Prototype support tools

37. 36SAS 2004 Publications Conferences Rouff, C., Vanderbilt, A., Hinchey, M. Truszkowski, W., Rash, J. Properties of a Formal Method for Prediction of Emergent Behaviors in Swarm-based Systems. 2nd IEEE International Conference on Software Engineering and Formal Methods. Beijing, China, 26-30 September, 2004. Rouff, C. Hinchey, M. and Truszkowski, W. Towards a Formal Approach to Swarm Technology. 8th International Conference on Space Operations (SpaceOps 2004). Montreal, Canada, May 17-21, 2004. Rouff, C., Vanderbilt, A., Truszkowski, W., Rash, J. and Hinchey, M. Verification of NASA Emergent Systems. First IFIP International Conference on Artificial Intelligence Applications and Innovations. August 23-26, 2004. Rouff, C., Vanderbilt, A., Truszkowski, W., Rash, J. and Hinchey, M. Verification of NASA Emergent Systems. Ninth International Conference on Engineering of Complex Computer Systems (ICECCS 2004), Florence, Italy, April 14-16, 2004.

38. 37SAS 2004 Publications Workshops Rouff, C., Vanderbilt, A., Hinchey, M., Truszkowski, W., and Rash, J. Verification of Emergent Behaviors in Swarm-based Systems. Workshop on Engineering of Autonomic Systems. In Proceedings of IEEE International Conference on Engineering of Computer Based Systems (ECBS 2004), Brno, Czech Republic, May 24-27, 2004. Rouff, C., Vanderbilt, A., Hinchey, M., Truszkowski, W., and Rash, J. Properties of a Formal Method for Modeling Emergence in Swarm-based Systems. In Proceedings of the Third NASA Goddard Workshop on Formal Approaches to Agent Based Systems. Springer, Lecture Notes in Artificial Intelligence. April 2004. Rouff, C., Truszkowski, W., Rash, J. and Hinchey, M. Formal Approaches to Intelligent Swarm Technology. IEEE/NASA Software Engineering Workshop. December 2003.