Adaptation and Maintenance of Multi-agent Systems using Reaction Rules
Xiao Liang · 2012
Software systems must be adaptable and maintainable in a business environment to remain useful.Business rules are central to business requirements and their adaptability and maintainability is the key to keep systems up to the changing environment over time.In this paper, we propose reaction rules for modelling interactive agent behaviour in Multi-Agent Systems.A rule-based knowledge model is sourced from the business requirements specification and becomes a knowledge base of driving agent behaviour dynamically, at runtime.Because the knowledge model is externalised from the running system itself, system behaviour can be adapted and maintained in an easier manner.The work is illustrated using the case of a rail track management system.