A new computing paradigm is currently on the spot: interaction based on series of actions. Most of autonomic agent systems (AASs) exploit this type of interaction to self-adjust their autonomous behaviors as a fundamental operational paradigm. At an interaction interface, actions evolve over time, hence series of actions occurs as a royal candidate for modeling, specifying, programming, and verifying AASs. For considering AASs, series of actions and adaptation relations; our formal approach consists, in particular, of categorical models and behaviors such that, firstly , AASs, series of actions and adaptation relations will categorically be modeled to provide algebraic frameworks for development of reasoning on their behaviors and, secondly, categorical behaviors of AASs, series of action and adaptation relations will be investigated and developed taking advantage of their categorical models.
TopIntroduction
For autonomic agent systems (AASs), autonomic computing is a generic property delineating capability to self-adjust their goal-driven computational behaviors without direct human interventions. Autonomic computing has been described as the set of concepts, technologies, and tools that enable AASs to become more self-managing. This potentiality is often related to possessing learning capabilities through analysis of past behaviors and interactions (Vinh, 2007, 2009a, 2009b, 2009c, 2009d, 2010; Vinh & Bowen, 2007, 2008). Autonomic computing has intensely been studied by various areas of engineering including agent systems, computational intelligent systems and human orientated systems (Vinh, 2010; Denko, Yang & Zhang, 2009; Jin & Liu, 2004; Pacheco, 2004; Witkowski & Stathis, 2004; Parashar & Hariri, 2006; Wang, 2007b; Ko, Gupta, & Jo, 2007; Yang & Liu, 2007; Butera, 2007; Calisti, Meer, & Strassner, 2008). With regard to AASs, autonomic computing (Wang, 2007b) and cognitive informatics (Wang, 2007a) have been set as two major pillars to support such systems. By the latest developments of autonomic computing (Vinh, 2009b; Denko, Yang & Zhang, 2009; Parashar & Hariri, 2006; Calisti, Meer, & Strassner, 2008) and cognitive informatics (Wang & Kinsner, 2006), AASs are now at a crucial point in their evolution, marked by research activities being booming (Vinh, 2009d; Topaloglu & Bayrak, 2008; S.A. DeLoach & Matson, 2008; F.E. Walter & Schweitzer, 2008; K. Zoethout & Molleman, 2008). AASs pose new challenges for the development and application of autonomic computing techniques, due to their special characteristics including: nondeterminism, context-awareness and goal- and inference-driven adaptability (Wang, 2007b).
AASs are agent systems, which implement autonomic computing mechanisms such as: nondeterminism, context-awareness and goal- and inference-driven adaptability. For autonomic computing techniques applicable to AASs, a new computing paradigm is currently on the spot: interaction based on series of actions. Most of AASs exploit this type of interaction to self-adjust their autonomous behaviors as a fundamental operational paradigm. At an interaction interface, actions evolve over time, hence series of actions occurs as a royal candidate for modeling, specifying, programming, and verifying AASs.
In this chapter, we focus on modeling AASs, series of action and adaptation relations, and then developing reasoning on their behaviors. Our formal approach consists mainly of categorical models and behaviors such that,
- •
Firstly, algebraic frameworks of AASs, series of action and adaptation relations will be constructed for development of reasoning on their behaviors using categorical language and,
- •
Secondly, categorical behaviors of AASs, series of action and adaptation relations will be considered taking advantage of their categorical models where the behavior-oriented notions will be formed for our approach.