Complex Systems Engineering Strategies.pdf
0 likes11 views
This document discusses complex systems engineering strategies. It describes properties of complex systems like emergence and self-organization. It outlines different behavior structures in complex systems from loose component collaboration to tight hierarchical control. Engineering strategies are proposed that take bottom-up self-organizing approaches, top-down traditional approaches, as well as middle-out and matchmaking approaches. Top-down, bottom-up, matchmaking and middle-out design methodologies are compared. The document concludes that complex systems require coordination strategies and controlled collaborative interactions to support useful emergent behaviors while avoiding chaos or rigidity.
Complex Systems Engineering Strategies.pdf
- 2. Complex Systems Properties • Emergent - Macroscopic dynamics and variables occur in the system, which is not easily predictable from local dynamics • Multiscale Interactive - The macroscopic and component-level behavior interact in a measurable way • Non-equilibrium – Large changes in the behavior of the system are possible under small perturbations • Metastable - Long-lived non-equilibrium states can arise and be maintained in interaction with the environment • Evolutionary Adaption- The system exhibits altered behavior in response to environmental changes • Self Organizing - Coordinated behavior can take place among components without centralized guidance
- 3. Behavior Structures in Complex Systems • Component = Basic element of functionality in the system. Intrinsic behavior under environment influences. • Collaboration = Interactions without macroscopic coordinators. Behavior influenced by peer-to-peer interactions. • Coordination = Interaction possibly managed by coordinators to support group goals. Behavior influenced by group dynamics. • Control = Interactions directed hierarchically to foster global goals. Behavior constrained by controller. • • All of these structures can be present in a system of systems and can be mixed and combined recursively
- 4. Components Behavior Structures Controller Components Collaboration Components Coordination Coordinators Controllers Coordinators Controllers Controllers Coordinators
- 5. Aspects and Examples of Behavior Components Collaboration Coordination Control No coupling Loose coupling Cooperative processes Tight Coupling Have dynamics and/or goals Individual dynamics and goals (Nash) Shared dynamics and goals (Pareto) Global dynamics and goals (Managed) Adapt to environment Individuals adapt to each other Individuals and coordinators adapt Controllers adapt Government, Societies Town Meetings, Interactions Representatives, Communities Authoritarian, Enterprises Molecules in Chemistry Gas Liquid Solid Computers in Network Peer-to-peer and Internet Extranet Intranet
- 6. Complex Systems Engineering Strategies • Bottom up – Self-organizing, Emergent collaboration and coordination from interactions. • Top down - Traditional systems engineering, Pre- defined coordination and interactions • Matchmaking – Coordination is based on matching and combining existing components to meet requirements • Middle Out – Coordination combines existing components and collaborations but also drives new requirements, collaborations and components Note: Alternate strategies can be used in different stages of engineering
- 7. Top Down Design (Control-based) Existing, Enhanced, Composite and possible new services and/or systems USERS Capability specifications (with priorities) Capability description Set of Services Needed Set of Requirements Requirements Analysis User Requirements and Derived Requirements
- 8. Bottom Up Design (Collaboration-based) Existing, Enhanced, Composite and possible new services and/or systems USERS Capability specifications and possible new capabilities Capability description (with cost of implementation) and possible new capabilities Capabilities available Set of Services Available Advertising User Interfaces
- 9. Matchmaking ( SOA Orchestration) Existing, Enhanced, Composite and possible new services and/or systems USERS Capability specifications (with priorities) Capability description (with cost of implementation) Set of Services Set of Requirements Matchmaking User Requirements and Derived Requirements
- 10. Middle Out Design (Coordination-based) Existing, Enhanced, Composite and possible new services and/or systems USERS Capability specifications (with priorities) and possible new capabilities Capability description (with cost of implementation) and possible new capabilities New capabilities available New capabilities needed Set of Services Set of Requirements Mediating User Requirements and Derived Requirements
- 11. Conclusions
- 12. Questions from 1995 • What are the basic laws of the scientific discipline of complex systems? • What are the generic principles for complex systems engineering? • Is it possible to build customizable generic tools for the modeling, simulation, and analysis of complex systems? • How can we maintain systems with constantly changing requirements? • Is there a management strategy for dealing with systems that are too complex for individuals or small groups to understand? • Are there unique characteristics of complex systems that are composed primarily of multiple intelligent entities, both human and non-human? • How can non-adaptable system elements be reengineered, and can adaptability be 'designed into' complex systems in the first place?
- 13. Final Thoughts • One of the fundamental change that complex systems bring to systems engineering is the need to develop federated coordination strategies as part of the engineering process • This coordination must be combined with controlled and collaborative interactions in an adaptive fashion that will support useful emergent behavior while avoid major negative outcomes caused by chaos or rigidity. • Implementing this new paradigm will require extensive research in multiple disciplines during the next decade. • Due to the broad fundamental impact of complex systems engineering, there should be a coordinated initiative to support research projects in this domain.