Experimental Setup and Parameters Sample Clauses

Experimental Setup and Parameters. For the experimental evaluation, we focus on faulty processes that will always try to make the system agree upon an incorrect value. The faulty processes have complete knowledge of the system including the correct value for each iteration. Our simulation uses two models for faulty processes. Model 1 uses a process that will always propose the incorrect value. Model 2 uses a process that looks at the percentage of its own weight to the weight of all processes and proposes the correct value if its percentage is below a threshold and the incorrect value otherwise. There are two types of non-faulty processes used. The first is an accurate non-faulty process that always proposes the correct value (d = 0.5, β = 1). The second type of non-faulty process chooses the correct value with probability β = 0.5 + d, where d ∈ [0, 0.5]. The Queen algorithm [BG89] is used for step 2 in the ABA algorithm and for all simulations, n = 41, f = 10 and ǫ = 0.1.