Mixed Initiative, Cooperative Self-adaptation for Systems of Systems
Abstract
The work will be conducted with the following tasks and milestones:Task 1: Extend our modeling and analysis framework built on the probabilistic modeling tool,PRISM, to support the extensions for analyzing coordination patterns in the context of delays,heterogeneity with respect to goals and capabilities, and different protocols. Evolvecommunication APIs for self-adaptive systems, implemented in Rainbow, to take advantage ofthese extensions.Task 2: Provide mechanisms that can adapt coordination patterns when environment conditionschange.Task 3: Extend our modeling and analysis of humans as effectors to also consider humans asdecision makers, as well as multiple human operators. Extend human models in Rainbow tohandle this multiplicity.Task 4: Develop a modeling and analysis framework for considering time when analyzinghuman involvement, considering both near- and long-term horizons.Task 5: Develop APIs and methods for integrating into the adaptive reasoning loop bothqualitative observations made by human operators, as well as the quantitative observations madeby sensors.Task 6: Identify an appropriate system of systems case study within the US Navy on which toexperiment and evaluate our techniques for mixed-initiative, coordinated self-adaptation.Schedule. We divide the schedule into 4 milestones as follows (time indicates how long after thestart of the project the milestone is due).Milestone 1 (6 months): Prototype modeling framework for incorporating delays and othertimings into our modeling framework. Application of this to communication delays andtimeliness of human involvement.Milestone 2 (12 months): Extended modeling framework to deal with multiple parties withmultiple goals. Application of this to coordination and multiple human operators. Design of APIto incorporate observations from humans into self-adaptation.Milestone 3 (24 months): Implementation of the design for human observations. Completion ofmodeling framework to consider long term horizons. Implementation of adaptive coordinationpatterns. Identify candidate Navy case study.Milestone 4 (36 months): Implementation of human models and new decision-making intoRainbow. Final evaluation on the case study
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 23, 2016
- Source ID
- N000141612961
Entities
People
- David Garlan
Organizations
- Massachusetts Institute of Technology
- Office of Naval Research
- United States Navy