I run the consensus algorithm, and for each round k, we record the norm of the disagreement vector(|(|δ(k)|)|>〖10〗^(-6)). We stop, at a predefined value|(|δ(k)|)|>〖10〗^(-6) and we call this “k_stop”. We have then a vector of these disagreement values as a function of the round number, which we can plot.
Now the next step is to see how the number of steps needed depends on the number of nodes, and on the average degrees, which in turn depends on the transmission range. Then, the step of the evaluation, we are following: We fix the network size, e.g., 100 nodes, and evaluate how the convergence time, i.e., the number of rounds needed to have the "local averages" equal to the known global average changes as we change the nodes transmission range. Thus we do as follows: -We take a given value of the transmission range, and generate 10 different networks, each with a given number of nodes , e.g., 100 nodes but in different random positions. For each network, measure the convergence time, and then take the average.
And then we had the two cases: 1) Iterations as a function of the number of nodes for given radius, and 2) Iterations as a function of the radius for given number of nodes Can you help me the explain how the number of steps needed depends on the number of nodes and depends on the transmission range and so on when the looking figure..