Arrow Research search

Author name cluster

D. N. Lam

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

2 papers
1 author row

Possible papers

2

AAAI Conference 2004 System Paper

Multi-Agent System Development: Design, Runtime, and Analysis

  • K. S. Barber
  • K. Fullam
  • N. Gujral
  • D. N. Lam
  • J. Park

Dynamic and unexpected events are defining characteristics of numerous application domains. These environments often require decision-makers to solve many problems with insufficient time and resources. To effectively assess options, decision-makers require situation analysis and decision-support tools that model the dynamism of these environments to make rapid, robust decisions. Autonomous agents and multi-agent systems satisfy these requirements for decision-making in dynamic environments. Developing a multi-agent system (MAS) is a challenging task, considering sophisticated agent interactions and uncertain environmental dynamics and domain requirements. This demonstration addresses the comprehensive development process for multi-agent systems; illustrating tools for the initial design of the agent system, the capabilities encoded in the individual agents, and analysis tools that enhance developer comprehension of system behavior.

PDF Details

AAAI Conference 2000 Conference Paper

Sensible Agents: Demonstration of Dynamic Adaptive Autonomy

  • K. S. Barber
  • D. C. Han
  • D. N. Lam
  • C. E. Martin

The analysis and design of large, complex systems mandates a formal methodology and supporting tools to assist system development teams throughout the system lifecycle. The multitude of personnel, the diversity of viewpoints, and the transient nature of personnel and technology in relation to the system lifecycle constrains the process by which 1) application domain requirements are acquired, analyzed and modeled, 2) a system architecture is derived from those requirements, 3) technology decisions are made and implementation progresses, and 4) the system is tested and maintained. A formal methodology for the entire lifecycle keeps team members coordinated and offers a mechanism to gauge progress. Large projects with many personnel responsible for making decisions require a formal process and automated support to assist team members in documenting their decisions. Traceability of decisions and documentation rationale is key to understanding the impact of decisions related to modeling, design, implementation, test, and maintenance. The SEPA effort proposes both a methodology and supporting tool suite (leveraging various knowledge representation and reasoning schemes) to facilitate development of object-oriented designs from evolving requirements. SEPA creates traceable, comprehensible, and extensible system design specifications based on requirements from system clients and domain experts. The funnel abstraction is chosen to represent the narrowing, refining, and structuring of user requirements into a system design. User inputs are refined by: (1) merging inputs from multiple sources, (2) distinguishing between inputs relating to system requirements and those relating to general domain knowledge, (3) constructing an object-oriented architecture, (4) mapping requirements to technology solutions, and (5) providing a framework for evaluating system design.

PDF Details
v2026.09.13