Artificial intelligence (AI) and robotics are rapidly developing areas in these days. AI is no longer just a theoretical area or some scientific experiment, it has became a technology often used in practice. Problem solving with single intelligent agents is now a widespread and working method, that's why with the increase of confidence in these kinds of systems the scope of application has constantly been expanding. However, for more complicated problems the usage of systems consisting of multiple agents is inevitable.
In order to take advantage of the multi-agent systems, it is necessary to establish the organization of the agents, and to define the protocols controlling the operation of the system. Such cooperation protocols play an important role in the regulation of agents' behaviour and communication, as they ensure that the system will work as expected. A group of these cooperation protocols is called the market-based auction protocols, which are very useful in the fields of task-sharing and resource reservation.
In my B.Sc. thesis I deal with specific types of auction protocols, which are capable of auctioning several items simultaneously. A variety of auction protocols have been developed for such cases, but the sequential auction stands out among them because of its beneficial properties. This auction protocol has already proven its qualities in practice. It is able to avoid the emerging NP-complete problems and use approximation procedures to finish the auction in linear time. As the practical part of my thesis, my goal is to create a software that models the sequential auction, and can be used for gathering practical experience about the operation of the protocol.