ρ | group of model parameters {a(t),b,σ}; |
\(c_{k}\) | number of parameters selected adaptively based on surrogate modeling; |
V | solution obtained by solving a full order model; |
\(c_{0}\) | Number of randomly selected parameter groups to initiate the algorithm; |
ε | error estimator; |
\(\varepsilon ^{k}\) | set comprised of error estimator values at kth iteration; |
\(e_{\mathrm{sg}}\) | set composed of all error estimator values at the kth iteration; |
ε̄ | surrogate error model; |
P̂ | parameter set used to obtain the optimal parameter group; |
\(\rho _{I}\) | optimal parameter group which maximizes the error estimator; |
e | relative error between a reduced and full model; |
V̄ | solution obtained using a reduced order model; |
V̂ | snapshot matrix; |
\(E_{p}\) | error set; |
ē | error model as an approximate error for the exact error e; |
\(e^{\mathrm{max}}_{\mathrm{tol}}\) | tolerance for the relative error, greedy iterations terminates if \(\bar{e} < e^{\mathrm{max}}_{\mathrm{tol}}\); |
\(\gamma _{e}\) | slope of the error model; |
τ̂ | intersection with the logarithmic axis log(y). |