Dynamic Programming and Greedy Heuristic in the Load Balancing Problem for the iFDAQ of the COMPASS Experiment at CERN

Abstract

Recently, a successful data taking in high energy physics experiments heavily relies on a stability of Data Acquisition System (DAQ). The paper deals with the Load Balancing (LB) problem of the intelligent, FPGA-based Data Acquisition System (iFDAQ) of the COMPASS experiment at CERN and presents two approaches –Dynamic Programming (DP) and Greedy Heuristic (GH) – applied in finding LB optimal solution. DP represents a pure mathematical approach in terms of a mathematical optimization and GH is a representative belonging to a class of heuristics. Therefore, the mutual comparison is worth investigating. In order to demonstrate the performance of the proposed DP and GH approaches, three LB test cases are examined and the numerical results are compared with other LB solution methods.