A CSP-Based Agent Modeling Framework for the Cougaar Agent-Based Architecture

Gračanin, Denis; Singh, H. Lally; Hinchey, Michael G.; Eltoweissy, Mohamed; Bohner, Shawn A.

doi:10.1109/ecbs.2005.6

Cited by 4 publications

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CSP has been used in other approaches for multi-agent systems. Examples include an approach that combines a CSP encoding of agent communications with a first-order logic framework [20]; a CSP framework for a Java-based "cognitive agent architecture" called Cougar [15], where the model is used to verify properties about the code generated from the Cougar system; and a timed CSP model of a multi-agent manufacturing system [31]. However, each of these approaches is (like ours) specific to its example application.…”

Section: Related Workmentioning

confidence: 99%

Formal Verification of a Map Merging Protocol in the Multi-Agent Programming Contest

Luckcuck¹,

C.²

2021

Preprint

View full text Add to dashboard Cite

Communication is a critical part of enabling multi-agent systems to cooperate. This means that applying formal methods to protocols governing communication within multi-agent systems provides useful confidence in its reliability. In this paper, we describe the formal verification of a complex communication protocol that coordinates agents merging maps of their environment. The protocol was used by the LFC team in the 2019 edition of the Multi-Agent Programming Contest (MAPC). Our specification of the protocol is written in Communicating Sequential Processes (CSP), which is a well-suited approach to specifying agent communication protocols due to its focus on concurrent communicating systems. We validate the specification's behaviour using scenarios where the correct behaviour is known, and verify that eventually all the maps have merged.

show abstract

Section: Related Workmentioning

confidence: 99%

Formal Verification of a Map Merging Protocol in the Multi-Agent Programming Contest

Luckcuck¹,

C.²

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Then the user maps the workflow of the intended system into its PIM and PSM using GDAM and GCAM components, respectively. An assembly approach is used, whereby the developers assembles the system and implementation models of the intended Cougaarbased system by choosing, configuring, and connecting various predefined GDAM and GCAM components [34]. Once completed, the models are fed into a transformer, which then parses through this assembled set of models to produce the actual software artifacts such as requirements, design, code, and test cases.…”

Section: Fig 4 Cmda Abstraction Layersmentioning

confidence: 99%

“…That said, there was nothing that indicated the use of other competing representations would be better or worst. We considered more formal representations like Z, B, and VDM++, and CSP [31,34]. OCL provided an effective means of representing domain and software constraints.…”

Section: Fig 4 Cmda Abstraction Layersmentioning

confidence: 99%

Model-based engineering for change-tolerant systems

Bohner

Ravichandar

Arthur

2007

Innovations Syst Softw Eng

View full text Add to dashboard Cite

Developing and evolving today's systems are often stymied by the sheer size and complexity of the capabilities being developed and integrated. At one end of the spectrum, we have sophisticated agent-based software with hundreds of thousands of collaborating nodes. These require modeling abstractions relevant to their complex workflow tasks as well as predictable transforms and mappings for the requisite elaborations and refinements that must be accomplished in composing these systems. At the other end of the spectrum, we have ever-increasing capabilities of reconfigurable hardware devices such as field-programmable gate arrays to support the emerging adaptability and flexibility needs of these systems. From a model-based engineering perspective, these challenges are very similar; both must move their abstraction and reuse levels up to meet growing productivity and quality objectives. Model-based engineering and software system variants such as the model-driven architecture (MDA) are increasingly being applied to systems development as the engineering community recognizes the benefits of managing complexity, separating key concerns, and automating transformations from high-level abstract requirements down through the implementation. However, there are challenges when it comes to establishing the correct boundaries for change-tolerant parts of the system. Capabilities engineering (CE) is a promising approach for defining long-lived components of a system to ensure some sense of change tolerance. For innovative initiatives such as the National Aeronautics and Space Administration (NASA)'s autonomous nanotechology swarms (ANTS), the development and subsequent evolution of such systems are of considerable importance as their missions involve complex, collaborative behaviors across distributed, reconfigurable satellites. In this paper, we investigate the intersection of these two technologies as they support the development of complex, change-tolerant systems. We present an effective approach for bounding computationally independent models so that, as they transition to the architecture, capabilitiesbased groupings of components are relevant to the changetolerant properties that must convey in the design solution space. The model-based engineering approach is validated via a fully functional prototype and verified by generating nontrivial multiagent systems and reusing components in subsequent systems. We build off of this research completed on the collaborative agent architecture, discuss the CE approach for the transition to architecture, and then examine how this will be applied in the reconfigurable computing community with the new National Science Foundation Center for High-Performance Reconfigurable Computing. Based on this work and extrapolating from similar efforts, the modelbased approach shows promise to reduce the complexities of software evolution and increase productivity-particularly as the model libraries are populated with canonical components.

show abstract

Formal Verification of a Map Merging Protocol in the Multi-agent Programming Contest

Luckcuck

Cardoso

2022

Engineering Multi-Agent Systems

View full text Add to dashboard Cite

A CSP-Based Agent Modeling Framework for the Cougaar Agent-Based Architecture

Cited by 4 publications

References 17 publications

Formal Verification of a Map Merging Protocol in the Multi-Agent Programming Contest

Formal Verification of a Map Merging Protocol in the Multi-Agent Programming Contest

Model-based engineering for change-tolerant systems

Formal Verification of a Map Merging Protocol in the Multi-agent Programming Contest

Contact Info

Product

Resources

About