Algorithms
Algorithms available within OSTRICH Development
Contents
1 Search Algorithms |
3 Multi-objective Calibrations |
|---|---|
2 Uncertainty Estimation |
4 Additional Algorithms |
1 Search Algorithms
Each algorithm has its own configuration group, wherein the user can specify the values for various algorithm control variables. Additional optional configuration variables and groups (i.e. Warm Start, Pre-Emption, Parameter Correction, a List of Initial Parameters, Math and Stats, and Line Search) may also be available for a given algorithm, as indicated in Table 1.
2 Uncertainty Estimation
Several of the search algorithms implemented in OSTRICH are designed to enumerate parameter probability distributions or behavioral parameter sets. Such algorithms are referred to as being “uncertainty-based” since they are not just concerned with identifying a single globally optimal parameter set. The configuration groups for these algorithms are described below.
DDS for Uncertainty Approximation |
Generalized Likelihood Uncertainty Estimation (GLUE) |
|---|---|
Metropolis-Hastings Markov Chain Monte Carlo (MCMC) |
Rejection Sampling |
3 Multi-objective Calibration
The algorithms described below seek to identify non-dominated solutions representing the tradeoff curve (i.e. pareto front) among conflicting objectives. These objectives can reflect a multi-criteria calibration exercise or a multi-objective optimization problem.
Pareto Archived DDS (PADDS) |
Asynchronous Parallel PADDS |
|---|---|
Simple Multi-Objective Optimization Test Heuristic (SMOOTH) |
4 Additional Algorithms
Mathematics and Statistics |
Line Search |
|---|---|
General-purpose Constrained Optimization Platform (GCOP) |