my doctoral research, completed at the universidad de castilla-la mancha (uclm) in 2024, focuses on the design of distributed control architectures that combine multi-agent systems (mas), bdi cognitive models, and rfid/epcis visibility frameworks to support real-time decision-making in complex logistics and airport environments.
the thesis proposes a three-layer architecture — bdi agent loop, mdp optimisation engine, and an is platform (erp + crm + expert system) — validated through the ciudad real central airport ground handling case study. the reward function R = 0.5·delay + 0.4·cost + 0.0·qos + 0.1·energy calibrates the system's operational priorities.
the mas-duo open-source library is the reference python implementation, built on pettingzoo aec, that translates all thesis concepts into executable code — from 4w state (what/where/when/why) to agent negotiation protocols and dynamic global policy updates.