These days more attention is given to the collective movement of agents, as it can be incorporated to different solutions in various application areas. These applications include agricultural (to monitor the spread of weeds and pests), mass event monitoring, rescue operations, border patrolling etc. The two of the most important aspects when implementing a flock, are the controlling mechanism and the communication model. In order to understand the flock algorithms better it is necessary to get a deep knowledge of the controlling and communication modells, that is why some of them have been introduced in my thesis.
Many problems arise when implementing a flocking solution, for example addressing the energyconsumption optimization issue or creating efficient algorithms for task allocation. In my thesis I have studied the implementation of a task allocation algorithm, which is optimized for flock tasks. I have examined different kind of scenarios in a simulation environment, which was created for this particular algorithm. I have observed the effects of the errors during communication (like packetlosses) on the performance of the flocking system. After the examination I have integrated a network module into the simulation environment, which allowed me to simulate the packet losses events, to check the effect of the packet loss on the simulated task allocation solution. Based on the results, I have proposed a solution, which can minimize the communication errors with the help of aditional acknowledgement messages.