Analysis of Passenger Movement at Birmingham International Airport Dr. Mario Gongora Center of Computation Intelligence De Montfort University Leicester, UK E-mail: mgongora@dmu.ac.uk Wasiq Ashfaq Center of Computation Intelligence De Montfort University Leicester, UK E-mail: wasiqashfaq@gmail.com Abstract— This paper presents a novel methodology for the analysis of the data at Birmingham Airport to provide effective and useful information about the dwell-time that passengers have between different points of their visit to the Airport. Birmingham Airport has sensors that anonymously count the number of people passing through crucial access routes, including boarding gates and security points. These sensors provide an enormous amount of crude data which contains valuable information reflecting the time people spend on dif- ferent parts of the premises, but extracting this information requires a complex processing of the data. The methodology presented in this work uses a Genetic Paradigm which is able to process that data using a compact and robust simulation model, so that the time spent by the visitors to the airport can be extracted from the raw data produced by the sensors. I. I NTRODUCTION Birmingham International Airport (BIA) is the UK’s fifth largest Airport, third largest for charter traffic and has the highest proportion of business traffic, second only to London Heathrow. Due to the key role BIA plays in the UK airline industry, it needs continuous upgrades of facilities and technology to meet the growing needs of the industry. Like any other business, in order to improve existing facilities or to introduce new services, the behaviour of the visitors needs to be understood. Their activities can help the concerned authorities plan different aspects of the airport management including positioning of shops, check- in counters, waiting lounge, toilets, payphones, fire escapes, security checks, etc. Furthermore, continuous studies can be used to identify the impact of different events on the behaviour of the customers. To understand the behaviour of the visitors at the airport, the management uses a variety of sensors to count the number of people passing through different points in both directions with an accuracy of up to 97%. These sensors provide the exact numbers of people entering or leaving a certain area at a certain time. Hence the data is anonymous and simple since there is no way of identifying visitors individually. However, the data holds a great deal of information regarding the behaviour of the visitors but needs proper processing and analysis to extract the desired information. The information extracted from the data will help BIA in better understanding the behaviour of the passengers and hence will enable them to provide better facilities and services. The methodology presented in this work uses a Genetic Paradigm which processes the sensor data using a compact and robust simulation model and extracts information regard- ing the time spent by the visitors in different areas of the airport. II. BACKGROUND AND LITERATURE REVIEW A. The problem description Airports are very busy places with thousands of visitors everyday. People go there for different reasons, for instance, arriving from other places, departing to other destinations, accompanying friends or family, working at the airports, etc. In this work, only the people who enter the departure lounge are dealt with, normally passengers and staff. For a passenger, there is only one entrance (through security) and many exits (to the airplanes), but for staff the same gate is used for entrance and exit. The counter data obtained has noise and although some of it can be filtered, some factors are impossible to prevent (e.g. people moving closely in groups). The sensors cannot individually track the customers, they only provide the count, therefore, the calculation of the dwell-time 1 has been a ‘grey-area’ for the management of BIA. Since there is no single correct way of calculating or validating the dwell-time, it is almost impossible to directly extract any further information e.g., the individual dwell- times, the dwell-times of different types of customers, etc. B. Airport simulations and other relevant work The steady growth of the airline industry demands contin- uous expansion of the facilities without affecting the smooth flow of operations. This makes airports an ideal application area for simulation. The environment is complex, stochastic and in a continuous state of change. The study and analysis of many different aspects of an airport has been of great interest throughout the world. This is evident from previous research work conducted at Schiphol Airport, Amsterdam [10], [27], [16], [18], [26], DFW Airport, USA [8], Washington International Airport, 1 The time (generally, on average) an individual passenger spends in a certain area