The Control of Multi-Agent Systems (MAS)
Abstract
Doubt continues that current theories and practices can produce agent or multiagent system (MAS) autonomy (Pynadath et al., 2001), imperiling future MAS missions (Lawless & Grayson, 2004). One possible reason is that from the view of methodological individualism (Nowak & Sigmund, 2004), reality is a stable phenomenon meaning that a single best view of reality is mathematically possible, a thesis that Benardete (2002) rejects. Alternatively, social reality is bistable (bistability occurs when a phenomena exists in one of two interdependent states; e.g., bistable interpretations refer to two stable, incommensurable interpretations of one data set, illustration, or phenomenon), and is best modeled by the quantum relations (Bohr, 1955). Given multiple interpretations of the same context or situation, to control an MAS, a group or decision maker can choose to avoid multiple interpretations as advised by Nash (1950) in order to converge mathematically to a consensus which we conclude is optimal to enact solutions for well-defined problems (wdp's), or to exploit these multiple interpretations which we conclude is the only way to solve ill-defined problems (idp's) (Lawless & Grayson, 2004).
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2004
- Accession Number
- ADA466002
Entities
People
- Margo Bergman
- Nick Feltovich
- W. F. Lawless
Organizations
- Paine College