Difference between revisions of "Alternate Run Modes:Multistart"
Beskahill2 (talk | contribs) (→Heading) |
(→Multistart) |
||
| (3 intermediate revisions by 2 users not shown) | |||
| Line 1: | Line 1: | ||
| + | === Multistart === | ||
| + | |||
This GSSHA alternate run mode incorporates a simple stochastic global optimization algorithm called Multistart, which involves both a global phase and a local phase. The Multistart method samples points from a uniform distribution over the feasible parameter space and starts a local search from each of the sample points. The local search algorithm is either the Levenberg–Marquardt (LM) method, or its efficiency enhancement, the secant LM (SLM) method. We recommend that one use the SLM method for the local searches. Uniform random sampling ensures global reliability of the method and the estimate for the global minimum is the smallest local minimum found. Multistart is inefficient in that each local minimum, particularly those in large regions of attraction in the parameter space, is generally found multiple times. Multistart naively stops after a user-specified number of iterations; however, more sophisticated stopping rules are possible. Prior to execution of this run mode, in addition to the activities that must be performed that uniformly apply to any of the four alternate GSSHA run modes (previously discussed in section 18.6), one must also prepare a file named ms.in (for “multistart input”). The contents of the file named ms.in are simply two integers whose values are specified on the first and second lines of the file wherein the first entry is a random number between 0 and 4,294,967,295, and the second integer value indicates the number of LM/SLM local searches that will be performed with initial points randomly sampled, uniformly, from feasible adjustable model parameter space. The required syntax to use this alternate GSSHA run mode for model calibration is: | This GSSHA alternate run mode incorporates a simple stochastic global optimization algorithm called Multistart, which involves both a global phase and a local phase. The Multistart method samples points from a uniform distribution over the feasible parameter space and starts a local search from each of the sample points. The local search algorithm is either the Levenberg–Marquardt (LM) method, or its efficiency enhancement, the secant LM (SLM) method. We recommend that one use the SLM method for the local searches. Uniform random sampling ensures global reliability of the method and the estimate for the global minimum is the smallest local minimum found. Multistart is inefficient in that each local minimum, particularly those in large regions of attraction in the parameter space, is generally found multiple times. Multistart naively stops after a user-specified number of iterations; however, more sophisticated stopping rules are possible. Prior to execution of this run mode, in addition to the activities that must be performed that uniformly apply to any of the four alternate GSSHA run modes (previously discussed in section 18.6), one must also prepare a file named ms.in (for “multistart input”). The contents of the file named ms.in are simply two integers whose values are specified on the first and second lines of the file wherein the first entry is a random number between 0 and 4,294,967,295, and the second integer value indicates the number of LM/SLM local searches that will be performed with initial points randomly sampled, uniformly, from feasible adjustable model parameter space. The required syntax to use this alternate GSSHA run mode for model calibration is: | ||
| + | |||
gssha –ms case.pst | gssha –ms case.pst | ||
| + | |||
where case.pst is the modified control file. The name of the primary model output file associated with this alternate GSSHA run mode is “slm_chl_ms1.rec”. Its contents include (1) an echo of the contents of the input file named “ms.in” and the case name of the control file employed, (2) a summary of the global phase of the multistart method consisting of a given row listing the names of the specified adjustable model parameters, the associated total objective function value named “Total”, and its various subcomponent names, followed by a set of rows, each one consisting of a sampled initial point, via uniform random sampling, and its related objective function value(s), and (3) a summary of the local phase of the method consisting of, for each LM/SLM local search performed, the initial and final parameter set and their associated objective function values, respectively. Upon execution of this alternate GSSHA run mode, another output file of potential interest is the record file whose name will be “case.rec”. Contents of the file named “case.rec” include more information summarizing the global and local phases associated with execution of the multistart method than what is provided in the primary output file previously described. Examples of these two output files are provided within the example problem files provided directly below. | where case.pst is the modified control file. The name of the primary model output file associated with this alternate GSSHA run mode is “slm_chl_ms1.rec”. Its contents include (1) an echo of the contents of the input file named “ms.in” and the case name of the control file employed, (2) a summary of the global phase of the multistart method consisting of a given row listing the names of the specified adjustable model parameters, the associated total objective function value named “Total”, and its various subcomponent names, followed by a set of rows, each one consisting of a sampled initial point, via uniform random sampling, and its related objective function value(s), and (3) a summary of the local phase of the method consisting of, for each LM/SLM local search performed, the initial and final parameter set and their associated objective function values, respectively. Upon execution of this alternate GSSHA run mode, another output file of potential interest is the record file whose name will be “case.rec”. Contents of the file named “case.rec” include more information summarizing the global and local phases associated with execution of the multistart method than what is provided in the primary output file previously described. Examples of these two output files are provided within the example problem files provided directly below. | ||
| − | [ | + | [[media:Test_prob_ms.zip|Example problem files (both input and output) associated with a Multistart supervised GSSHA alternate run mode model calibration run]], which are supplied for use as a go by |
<noinclude> | <noinclude> | ||
{{Nav|Nav18}} | {{Nav|Nav18}} | ||
</noinclude> | </noinclude> | ||
Latest revision as of 17:48, 22 July 2013
Multistart
This GSSHA alternate run mode incorporates a simple stochastic global optimization algorithm called Multistart, which involves both a global phase and a local phase. The Multistart method samples points from a uniform distribution over the feasible parameter space and starts a local search from each of the sample points. The local search algorithm is either the Levenberg–Marquardt (LM) method, or its efficiency enhancement, the secant LM (SLM) method. We recommend that one use the SLM method for the local searches. Uniform random sampling ensures global reliability of the method and the estimate for the global minimum is the smallest local minimum found. Multistart is inefficient in that each local minimum, particularly those in large regions of attraction in the parameter space, is generally found multiple times. Multistart naively stops after a user-specified number of iterations; however, more sophisticated stopping rules are possible. Prior to execution of this run mode, in addition to the activities that must be performed that uniformly apply to any of the four alternate GSSHA run modes (previously discussed in section 18.6), one must also prepare a file named ms.in (for “multistart input”). The contents of the file named ms.in are simply two integers whose values are specified on the first and second lines of the file wherein the first entry is a random number between 0 and 4,294,967,295, and the second integer value indicates the number of LM/SLM local searches that will be performed with initial points randomly sampled, uniformly, from feasible adjustable model parameter space. The required syntax to use this alternate GSSHA run mode for model calibration is:
gssha –ms case.pst
where case.pst is the modified control file. The name of the primary model output file associated with this alternate GSSHA run mode is “slm_chl_ms1.rec”. Its contents include (1) an echo of the contents of the input file named “ms.in” and the case name of the control file employed, (2) a summary of the global phase of the multistart method consisting of a given row listing the names of the specified adjustable model parameters, the associated total objective function value named “Total”, and its various subcomponent names, followed by a set of rows, each one consisting of a sampled initial point, via uniform random sampling, and its related objective function value(s), and (3) a summary of the local phase of the method consisting of, for each LM/SLM local search performed, the initial and final parameter set and their associated objective function values, respectively. Upon execution of this alternate GSSHA run mode, another output file of potential interest is the record file whose name will be “case.rec”. Contents of the file named “case.rec” include more information summarizing the global and local phases associated with execution of the multistart method than what is provided in the primary output file previously described. Examples of these two output files are provided within the example problem files provided directly below.
Example problem files (both input and output) associated with a Multistart supervised GSSHA alternate run mode model calibration run, which are supplied for use as a go by
GSSHA User's Manual
- 18 Alternate Run Modes
- 18.1 MPI and OpenMP Parallelization
- 18.2 Simulation Setup for Alternate Run Modes
- 18.3 Batch Mode Runs
- 18.4 Automated Calibration with Shuffled Complex Evolution
- 18.5 Monte Carlo Runs
- 18.6 ERDC Automated Model Calibration Software
- 18.6.1 Efficient Local Search
- 18.6.2 Multistart
- 18.6.3 Trajectory Repulsion
- 18.6.4 Effective and Efficient Stochastic Global Optimization
- 18.7 Inset Models
