Adaptive decision making in agent-based simulation

Abstract

In the context of adaptive and autonomous decision making, a computational model is needed to formalize and implement a practical goal deliberation mechanism that determines how goals can be evaluated, adopted, or rejected. Such a model is expected to lead to self-explanatory decision making, which is essential for understanding and influencing the behavior of autonomous agents and to pave the way to generate desirable macro-level behavior in self-adaptive, self-organizing systems. This article introduces an adaptive decision making architecture for agent-based simulation by promoting deliberative coherence and its extensions for decision making under uncertainty. To this end, a deliberative coherence-driven agent model is presented, and then is extended with run-time monitoring mechanisms. The architecture enables us to evaluate goals and competing tasks to facilitate the selection of the most coherent tasks with respect to a given goal, whereas the successful completion of those tasks contributes to the achievement of mission objectives for systems in shifting, ill-defined, and uncertain environments.

Rapid prototyping of cognitive agent simulations using C-BML transformations

“…DeCoMeta defines the abstract syntax of DSL for the deliberative coherence theory that is implemented as DeCo library in DeCoAgent. 20 DeCoMeta facilitates defining a PIM (i.e., DeCoModel) and contains the structural modeling elements of the concrete syntax (i.e., the connectionist network) to enable construction of the connectionist network, called the DeCo network (DeCoNet). The metamodel is presented in the following section.…”

“…In our earlier work, 20 we put TDC in an agent-based simulation context, focusing the implementation perspective of TDC with a systems engineering view point, to show how an adaptive decision making can be realized. As a result, the deliberative coherence driven agent (DeCoAgent) is proposed as a coherence-driven agent that has adaptive decision-making ability based on the TDC computations.…”

“…As a result, the deliberative coherence driven agent (DeCoAgent) is proposed as a coherence-driven agent that has adaptive decision-making ability based on the TDC computations. 20 A practitioner may implement DeCoAgents employing DeCoAgent library, 21 which is available as open source. 22 The DeCoAgent library is implemented as an extension to Jadex, 23 which is an agent development and execution environment.…”

Remote Operation SW for USV: Part I. Integrated Mission Planning System

“…The integrated mission planning process of USV is an intelligent system that has high-level decision-making functions such as mission planning, path planning, and equipment operation planning to smoothly perform a given mission in an unknown marine environment [10]. Depending on the given mission and envi-…”

“…This shows a multi-agent hierarchical structure combining the above-described unit agents of different functions. In Figure 5, this process introduced Norman's seven-step model of human behavior processing in cognitive science [10] [11]. This model is fundamentally same with artificial intelligence or humanoid that approach three basic elements:…”