Algorithms for processing of large data sets using distributed architectures and load balancing

Modern experiments in high energy physics impose great demands on the reliability, the efficiency, and the data rate of Data Acquisition System (DAQ). This thesis focuses on the stability of the intelligent, FPGA-based Data Acquisition System (iFDAQ) of the COMPASS experiment at CERN. The iFDAQ utilizing a hardware event builder is designed to be able to readout data at the maximum rate of the experiment. The iFDAQ stability improvement consists of several various tasks. Firstly, the new communication library DIALOG for the Inter-Process Communication (IPC) is presented. Secondly, the DAQ Debugger is developed to help with the iFDAQ error detection. Then, the stable iFDAQ gives an opportunity to implement the iFDAQ continuously running mode running 24/7 without any stops. Finally, Load Balancing (LB) of the iFDAQ is solved using Dynamic Programming (DP), Greedy Heuristic (GH), Integer Linear Programming (ILP), Genetic Algorithm (GA) and Reinforcement Learning (RL).