Component-based Agent Research Articles

A self-switching multi-strategic pedagogical agent

Multi-Strategic Pedagogical Systems (MSPSs) integrate multiple Pedagogical Strategies (PSs) and a Switching Logic (SL). They trigger the PS appropriate regarding the Student State (StS). However, the hard-coding and the monolithic integration of multiple PSs and a SL highly increase the complexity of the MSPSs. To overcome this deficiency, we recourse to the Self-* and the Component-Based Agents design approaches to engineer a Self-Switching Multi-Strategic Pedagogical Agent (Self-S MSPA). It is designed as two distinct interacting Manager (MrSS) and Managed (MdSS) Sub-Systems. The MdSS implements the PSs. The MrSS implements the SL in the form of an Autonomic Control Loop (ACL) to (re)configure the MdSS. To experiment the usefulness of the proposed model, a framework is constructed using SAFRAN.

Augmenting agent computational environments with quantitative reasoning modules and customisable bridge rules

There are many examples where large amounts of data might be potentially accessible to an agent, but the agent is constrained by the available budget since access to knowledge-bases is subject to fees. Also, there are several activities that an agent might plan and perform on the web where one or more stages imply the payment of fees. For instance, consider the issue of buying resources in a cloud computing context where the objective of the agent is to obtain the best possible configuration of a certain application within given budget constraints. In this paper, we consider the software-engineering problem of how to practically empower agents with the capability to perform budget-constrained reasoning in a uniform and principled way. To this aim, we enhance the ACE component-based agent architecture by means of a device for practical and computationally affordable quantitative reasoning, whose results actually determine one or more courses of agent's actions, also according to policies/preferences. We further enhance the ACE framework by making the agent-components interaction mechanism parametric with respect to the actual modules that an agent may dynamically decide to exploit. We discuss the proposed extensions on a realistic case-study.

Dynamics and control in component-based agent models

Dynamics are an important aspect of agent models. Control of dynamics requires specific methods of specification that have their own specific semantics. This paper addresses specification and semantics of dynamics and control in component-based agent models. Specification is based on a dedicated formal design specification language for agent models. Semantics of the dynamics are defined using temporal traces with composite states. It is shown in what manner control aspects can be specified, and what their semantics is in terms of the temporal traces. An agent model for controlled diagnostic reasoning processes is used to illustrate the approach. © 2002 Wiley Periodicals, Inc.