Enhancement of Airport Collaborative Decision Making through Applying Agent System with Matching Theory Antonio C. de Arruda Junior TransLab, CIC, CP 4466 University of Brasilia – UnB Brasilia-DF, 70910-900, Brazil jnarrd@yahoo.com.br Li Weigang TransLab, CIC, CP4466 University of Brasilia – UnB Brasilia-DF, 70910-900, Brazil weigang@unb.br Kamila B. Nogueira Ministry of Education Esplanada dos Ministérios, Brasília-DF, 70047-900, Brazil kamila.b.nogueira@gmail.com ABSTRACT The Collaborative Decision Making (CDM) paradigm attempts to improve the exchange of information among the various stakeholders involved in Air Traffic Management (ATM). It is aimed at efficient decision making in airport management. Although the processes of CDM are considered mature and well accepted, in many cases its focus is on the information sharing and is still not able to simultaneously involve essential agents such as Air Traffic Control (ATC) agency, airlines, and airport managers in the decision making. This study uses the matching approach of Game Theory to construct a two-sided matching market model for slot allocation in the Compression step while taking into account Ground Delay Programs (GDP). Our proposed model, Deferred Acceptance CDM (DA-CDM), assigns each flight to each slot through a "one-to-one" relationship, respecting the preferences of each allocation, leading to a stable result. It is applied to evaluate the classic CDM and Airport CDM processes with a group of analytics data. Our results show that the new allocation mechanism provides a stable and satisfactory matching of the flights with the slots in A-CDM procedure. Categories and Subject Descriptors I.2.11 [Artificial Intelligence]: Distributed Artificial Intelligence – Intelligent agents, Multiagent systems; I.6.5 Computing Methodologies, Simulation and Modeling, Model Development. General Terms Algorithms, Management, Design, Theory Keywords Multiagent Systems, Collaborative Decision Making, Ground Delay Program, Matching Theory. 1. INTRODUCTION Over the last few years, the increasing global demand for air transportation has greatly increased the complexity of the air traffic management scenario [5]. This situation enforces new integration challenges faced by several stakeholders, such as regulation agents, airlines, airport management companies, traffic managers, flight crew, passengers, and aeronautical system’s manufacturers, among others [17]. Proceedings of the 8th International Workshop on Agents in Traffic and Transportation (ATT-2014), Vizzari, Kluegl, Vokrinek (eds.), held at the 13th Int. Conf. on Autonomous Agents and Multiagent Systems (AAMAS 2014), May 5-6, 2014, Paris, France. Some processes, such as those aimed at reducing congestion in specific locations in the air scenario, involves the definition of delays for aircraft on ground and are carried through the Ground Delay Program (GDP). This process, based on Collaborative Decision Making (CDM) concepts, brings the need of reallocating aircraft from the scheduled slots originally established for the runways of the affected airports [27]. Besides its simplicity of concepts, the current CDM model involves a limited number of entities in the decision-making process[6]. When using traditional CDM model and considering the existence of distinct interests on delays applied to aircraft, it is a difficult task to get the satisfaction of all stakeholders who affect and are affected by delays generated by a GDP [23]. In this context, the matching approach of Game Theory can be used to construct the model of markets with the satisfactory results regarding the dispute for resources. By this approach, the preferences of all participants in that market are taken into account [25]. Regardless of the application area, a market can be modeled in order to obtain results that account for the different goals multiple agents, such as students, schools, doctors, hospitals, patients, passengers, airlines, and airports, among others. Moreover, the modeling constraints on organ donation markets in the 2000s allowed the correct treatment of a wide variety of features in more complex scenarios [21, 22]. In situations involving the departure coordination, traffic, and arrival of multiple flights through Air Traffic Management (ATM), mathematicians, economists, engineers, computer scientists, and researchers from various fields have developed Artificial Intelligence, multi-agent systems, and models based on Game Theory, among others. These models are applied in domains that involve problems of coordination and competition for resources [1, 3, 9, 28, 29]. However, most of the studies dealing with problems regarding GDP take into account only the interests of traffic control institutions and airlines. The limitation of these works based on the classic CDM model might lead to a limited level of satisfaction among other agents in the CDM process, and, consequently, the results of the process may not be stable [23]. In Brazil, this fact can be verified by the current situation, in which several concessionaires formed by private companies are entering the market to manage the major airports of the country [15]. Several projects which aim to improve the quality of services and airport infrastructure and to enlarge the supply of air transport to Brazilian population, currently handle billions of reais (Brazil’s currency) with the duration of 20 to 30 years, depending on the granting rules. Although the role of the airport operators is of crucial significance, the ATM process currently only accounts for the