際際滷

際際滷Share a Scribd company logo

Overview of a Self-Adaptive Workflow System

Download as PPT, PDF
Luca Sabatucci
Luca Sabatucci

The document discusses enabling self-adaptive workflows through a multi-agent system approach. It proposes representing business processes as goals that can be decomposed and addressed by autonomous software agents. The agents use a BDI architecture and self-organize to form teams capable of addressing goals. This allows workflows to flexibly adapt to changes in context by relaxing constraints or forming new teams if tasks fail. The system aims to make business processes more robust by enabling goals to be addressed through different "hows" based on an agent's capabilities and the execution context.

1 of 21
Download to read offline
Consiglio Nazionale delle Ricerche Istituto di Calcolo e Reti ad Alte Prestazioni Towards Self-Adaptation and Evolution in Business Process Luca Sabatucci 15 Ottobre 2013
Overview of a Self-Adaptive Workflow System
BPMN Commercial workflow engines follow the plan as in the blueprint (BPMN, BPEL) The execution model is based on the petri-net model: token and activation
Self-Adaptive Workflows In current workflow systems, a unexpected situation generates a fail to the standard plan, alternative plans can not be autonomously considered In order to enable self-adaptation, rigid constraints of a workflow must be relaxed plans should be searched in a broader solution space.
Our Approach
FROM BPMN to Goals
Decoupling WHAT and HOW We desire that BPMN is used for describing the result to address, not how to address it We use a Goal-Oriented approach GoalSPEC: a language to describe business goals to delegate to the system
Goal in GoalSpec Goal TRIGGERING CONDITION The state of the world that must hold because the goal becomes active ACTOR LIST Who is responsible FINAL STATE The state of the world that must be true for considering the goal addressed In oorderto des In rder to des ibe cr cr Trigggercco dit ibe Tri ger onndit n io n io andd an FFinalstaate inal st te weeuusean oo to w se an nnto gica lo gicalapppro h lo l ap roac h ac (WHEN MESSAGE book_request(Book) RECEIVED FROM THE Client ROLE AND WHEN available(Book) ) THE Clerk ROLE SHALL ADDRESS book_checkout(Book,Client) (BookManagment example) Each goal in GoalSPEC Each goal in GoalSPEC defines aa defines ition, desireddStateeTrans ition, desire Stat Trans G:TC -> FS G:TC -> FS
Extracting GoalSPEC from BPMN The workflow execution may be seen as a finite set of state transitions from the start event to the end event. Where: each FlowNode generates a single step of the transition We identify intermediate transitions
Examples Every FlowNode potentially impacts the state of the workflow WHEN completed(client_credential) THE user role SHALL ADDRESS ( done(booking) AND sent(receipt,client) ) OR error(booking) AFTER 1 hour SINCE WHEN done(login) OR WHEN user_number > MAX_USERS OR WHEN thrown(stopping_signal) THE user role SHALL ADRESS done (logout)
THE Agent ARCHITECTURE
Why Software Agents Agents encapsulate autonomy and proactivity. Agents ground on the classic AI loopmodel: sense reason act BDI agents represent a good abstraction for self-awareness and contextawareness MAS are a flexible and powerful method for distributed reasoning
The BDI Architecture (mental states) Agent1 Agent2 AgentN (plans) (act and perceive) Environment 13
Context-Awareness and SelfAwareness Proactive contextual matching of the behavior (how) with expected results (what) The agent must be able to reason on evolving business goals on the execution context on its capabilities
The Self-Aware Agent SYSTEM GOALS Self-Aware agents knows: system goals their own execution state their capabilities and how capabilities can be used (means-end reasoning) CAPABILITIES Agent1 Agent2 AgentN (plans) (UI) Capability encapsulat es the ability to manipulat e resources, to call web services or to interact with humans Environment System goals derive from business analysis, when some business goal is delegated to the system t) en tm mi m (co (UI) Resources (databases, ) Business Analyst Web Service Web Service Workflow User 15
Self-Organization Algorithm Goal commitment is a social activity Each goal prescribes a state transition (TC -> FS) Agents may own capabilities for addressing sub-transitions (sinput -> soutput) A Solution is a decomposition of the main transition TC -> FS into a set of sub-transitions A Team prescribes a collaboration among many agents, and it is regulated by contracts and rewards.
A solution is a set of potential contracts for addressing sub-goals: contract( potential commitment, curriculum, request income) TThealalgori m he gorith thmfor distriribute an forsearching gsolutitions is dist butedd an re searchin solu ons is dd recursiv cursivee system-goal goal-set {TC | FS} sub-goal G1..G5 G1 G2 {TC | IS1} AGENTS G3 {IS1 | IS2} {IS2 | IS3} G5 {IS3 | IS4} I may commit to G3 A1 ch Agents tries sto mat ch Agents trie to mat goal s their rcapacicitieswith goal thei capa tie with transitionnTC -> FS transitio TC -> FS de IF Agents alalsodecicide IF Agents so de the PARTICIPATEEto the PARTICIPAT to H solutionnanddWHIC H solutio an WHIC PART play in it, PART play in it, r depending gon thei r dependin on thei workinngqueuee worki g queu G4 AN Aj Recursionnstops w Recursio stops w n no he he solutionnisisdiscover n no solutio discoveeddor in re or in theetrivivialcase of nu th tr ial case of nu ll ll transition transition contract( {IS2 | IS3}, my_curriculum, 1) backward-goal decompose and recursively search sub-goals {TS | IS2} A1 A2 {IS4 | FS} forward-goal {IS3 | FS} AN A1 A2 AN ay be Many ysolutions m ay be Man solutions m discovered. . discovered alu ed ev ated Each solutionnisisev aluat Each solutio according gto: : accordin to ts completeness,s,agen ts ) completenes agen cost ) tal st reputationnanddto tal co reputatio an to
Conclusions: The new Lifecycle of Business Process Business Analyst models Theebusiness sanalyst Th busines analyst itional l continues sto useetrad itiona continue to us tradhis odel his instruments to m odel instruments to m processes s processe revises The system automatically translates BPMN into goals Every yagent in th Ever agent in th syst ee em autonomously de system autonomously de des ci whennto commit cides whe to committo so me to some goalal go GoalSPEC isis a GoalSPEC a language for language for expressing ggoals expressin goals that tgrounds son tha ground on ontology yandd ontolog an defines sa agoalalas the define go as the tuple tuple (actor,r, (acto trigger-condition, trigger-condition, final-state) ) final-state GoalSPEC injection worker Business Expert interacts commits to analyzes analyzes Running MAS analyzes commits to analyzes 18 Environment perturbations
Conclusions: The components of the MAS Solution Agents are specialized: every agent owns its own capacities. User capacities are used to interact with humans and monitor their activity Service capacities are used to manipulate the environment. Agent are peers: there is not a pre-established organization. They organize themselves in teams every time a new workflow starts The self-org algorithm considers many criteria (experience, cost, trust) The team is the candidate group of agents for addressing the workflow, in a given context Anyway, the context may change for some reason during the execution: In case of starvation the team tries to relax some constraints If some task fails, the team is dismissed and a new team is formed If no alternative team can be found, the analyst is informed of failure Commitment: when an agent is involved in a team, it tries to address its responsibilities at best of its possibilities It waits for triggering conditions hold It selects and executes the proper capacity or composition of capacities It checks that result is the expected final state Trust and Reward:: agents that successfully complete their task gain reputation and they increase their chance to be selected again.
Future Works Agents Learns by Experience to improve owned capabilities Studying to acquire new capabilities Coupling Goals and Norms in GoalSPEC Open Systems and Clouds
Thanks for your Attention Luca Sabatucci sabatucci@pa.icar.cnr.it Consiglio Nazionale delle Ricerche Istituto di Calcolo e Reti ad Alte Prestazioni

More Related Content

Overview of a Self-Adaptive Workflow System

  • 1. Consiglio Nazionale delle Ricerche Istituto di Calcolo e Reti ad Alte Prestazioni Towards Self-Adaptation and Evolution in Business Process Luca Sabatucci 15 Ottobre 2013
  • 3. BPMN Commercial workflow engines follow the plan as in the blueprint (BPMN, BPEL) The execution model is based on the petri-net model: token and activation
  • 4. Self-Adaptive Workflows In current workflow systems, a unexpected situation generates a fail to the standard plan, alternative plans can not be autonomously considered In order to enable self-adaptation, rigid constraints of a workflow must be relaxed plans should be searched in a broader solution space.
  • 6. FROM BPMN to Goals
  • 7. Decoupling WHAT and HOW We desire that BPMN is used for describing the result to address, not how to address it We use a Goal-Oriented approach GoalSPEC: a language to describe business goals to delegate to the system
  • 8. Goal in GoalSpec Goal TRIGGERING CONDITION The state of the world that must hold because the goal becomes active ACTOR LIST Who is responsible FINAL STATE The state of the world that must be true for considering the goal addressed In oorderto des In rder to des ibe cr cr Trigggercco dit ibe Tri ger onndit n io n io andd an FFinalstaate inal st te weeuusean oo to w se an nnto gica lo gicalapppro h lo l ap roac h ac (WHEN MESSAGE book_request(Book) RECEIVED FROM THE Client ROLE AND WHEN available(Book) ) THE Clerk ROLE SHALL ADDRESS book_checkout(Book,Client) (BookManagment example) Each goal in GoalSPEC Each goal in GoalSPEC defines aa defines ition, desireddStateeTrans ition, desire Stat Trans G:TC -> FS G:TC -> FS
  • 9. Extracting GoalSPEC from BPMN The workflow execution may be seen as a finite set of state transitions from the start event to the end event. Where: each FlowNode generates a single step of the transition We identify intermediate transitions
  • 10. Examples Every FlowNode potentially impacts the state of the workflow WHEN completed(client_credential) THE user role SHALL ADDRESS ( done(booking) AND sent(receipt,client) ) OR error(booking) AFTER 1 hour SINCE WHEN done(login) OR WHEN user_number > MAX_USERS OR WHEN thrown(stopping_signal) THE user role SHALL ADRESS done (logout)
  • 12. Why Software Agents Agents encapsulate autonomy and proactivity. Agents ground on the classic AI loopmodel: sense reason act BDI agents represent a good abstraction for self-awareness and contextawareness MAS are a flexible and powerful method for distributed reasoning
  • 13. The BDI Architecture (mental states) Agent1 Agent2 AgentN (plans) (act and perceive) Environment 13
  • 14. Context-Awareness and SelfAwareness Proactive contextual matching of the behavior (how) with expected results (what) The agent must be able to reason on evolving business goals on the execution context on its capabilities
  • 15. The Self-Aware Agent SYSTEM GOALS Self-Aware agents knows: system goals their own execution state their capabilities and how capabilities can be used (means-end reasoning) CAPABILITIES Agent1 Agent2 AgentN (plans) (UI) Capability encapsulat es the ability to manipulat e resources, to call web services or to interact with humans Environment System goals derive from business analysis, when some business goal is delegated to the system t) en tm mi m (co (UI) Resources (databases, ) Business Analyst Web Service Web Service Workflow User 15
  • 16. Self-Organization Algorithm Goal commitment is a social activity Each goal prescribes a state transition (TC -> FS) Agents may own capabilities for addressing sub-transitions (sinput -> soutput) A Solution is a decomposition of the main transition TC -> FS into a set of sub-transitions A Team prescribes a collaboration among many agents, and it is regulated by contracts and rewards.
  • 17. A solution is a set of potential contracts for addressing sub-goals: contract( potential commitment, curriculum, request income) TThealalgori m he gorith thmfor distriribute an forsearching gsolutitions is dist butedd an re searchin solu ons is dd recursiv cursivee system-goal goal-set {TC | FS} sub-goal G1..G5 G1 G2 {TC | IS1} AGENTS G3 {IS1 | IS2} {IS2 | IS3} G5 {IS3 | IS4} I may commit to G3 A1 ch Agents tries sto mat ch Agents trie to mat goal s their rcapacicitieswith goal thei capa tie with transitionnTC -> FS transitio TC -> FS de IF Agents alalsodecicide IF Agents so de the PARTICIPATEEto the PARTICIPAT to H solutionnanddWHIC H solutio an WHIC PART play in it, PART play in it, r depending gon thei r dependin on thei workinngqueuee worki g queu G4 AN Aj Recursionnstops w Recursio stops w n no he he solutionnisisdiscover n no solutio discoveeddor in re or in theetrivivialcase of nu th tr ial case of nu ll ll transition transition contract( {IS2 | IS3}, my_curriculum, 1) backward-goal decompose and recursively search sub-goals {TS | IS2} A1 A2 {IS4 | FS} forward-goal {IS3 | FS} AN A1 A2 AN ay be Many ysolutions m ay be Man solutions m discovered. . discovered alu ed ev ated Each solutionnisisev aluat Each solutio according gto: : accordin to ts completeness,s,agen ts ) completenes agen cost ) tal st reputationnanddto tal co reputatio an to
  • 18. Conclusions: The new Lifecycle of Business Process Business Analyst models Theebusiness sanalyst Th busines analyst itional l continues sto useetrad itiona continue to us tradhis odel his instruments to m odel instruments to m processes s processe revises The system automatically translates BPMN into goals Every yagent in th Ever agent in th syst ee em autonomously de system autonomously de des ci whennto commit cides whe to committo so me to some goalal go GoalSPEC isis a GoalSPEC a language for language for expressing ggoals expressin goals that tgrounds son tha ground on ontology yandd ontolog an defines sa agoalalas the define go as the tuple tuple (actor,r, (acto trigger-condition, trigger-condition, final-state) ) final-state GoalSPEC injection worker Business Expert interacts commits to analyzes analyzes Running MAS analyzes commits to analyzes 18 Environment perturbations
  • 19. Conclusions: The components of the MAS Solution Agents are specialized: every agent owns its own capacities. User capacities are used to interact with humans and monitor their activity Service capacities are used to manipulate the environment. Agent are peers: there is not a pre-established organization. They organize themselves in teams every time a new workflow starts The self-org algorithm considers many criteria (experience, cost, trust) The team is the candidate group of agents for addressing the workflow, in a given context Anyway, the context may change for some reason during the execution: In case of starvation the team tries to relax some constraints If some task fails, the team is dismissed and a new team is formed If no alternative team can be found, the analyst is informed of failure Commitment: when an agent is involved in a team, it tries to address its responsibilities at best of its possibilities It waits for triggering conditions hold It selects and executes the proper capacity or composition of capacities It checks that result is the expected final state Trust and Reward:: agents that successfully complete their task gain reputation and they increase their chance to be selected again.
  • 20. Future Works Agents Learns by Experience to improve owned capabilities Studying to acquire new capabilities Coupling Goals and Norms in GoalSPEC Open Systems and Clouds
  • 21. Thanks for your Attention Luca Sabatucci sabatucci@pa.icar.cnr.it Consiglio Nazionale delle Ricerche Istituto di Calcolo e Reti ad Alte Prestazioni