BlueM.Opt: Difference between revisions

Revision as of 07:01, 30 June 2009

Simulation-based optimization

Screenshot

Description

BlueM.Opt is an optimization framework that can be coupled with an arbitrary simulation software (only requirement: input data and results are to be stored in ASCII format). The optimization parameters, objective functions and (optionally) boundary conditions are defined in a flexible manner.

Optimization results are stored in a database.

BlueM.Opt integrates a graphing feature for displaying the optimization progress and results. Optimization results can also be analyzed in detail.

Where possible, BlueM.Opt utilizes multithreading in order to evaluate multiple parameter sets simultaneously.

List of available methods (optimization algorithms):

PES: Parametric Evolution Strategy
CES: Combinatorial Evolution Strategy
HYBRID: Combination of PES and CES
Hooke & Jeeves: Hillclimbing Algorithm
MetaEvo: multicritera, hybrid optimization algorithm
DDS: Dynamically Dimensioned Search

another method is

SensiPlot: Sensitivity analysis (no optimization)
Traveling Salesman

List of currently implemented simulation models:

BlueM.Sim
SMUSI
S:CAN
SWMM
Test problems

Downloads

BlueM.Opt Downloads

Application

Instructions for the Application of BlueM.Opt (PDF-Export)
Documentation of the input files
Documentation of the output files

Development

BlueM.Opt Development
- Disambiguation
- Documentation of the Codes
- Bugs management: Bugzilla
- Compilation

Internal

Documentation of the SVN Repository

@@ Line 4: / Line 4: @@
 [[Bild:EVO Box screenshot.png|thumb|Screenshot]]
 ==Description==
-BlueM.Opt is an optimization framework that allows an arbitrary simulation software to combine (only requirement: Input data and results are to be stored in ASCII format) with an optimization. The optimized parameters, objective functions and (optionally) boundary can be easily determined.
+BlueM.Opt is an optimization framework that can be coupled with an arbitrary simulation software (only requirement: input data and results are to be stored in ASCII format). The optimization parameters, objective functions and (optionally) boundary conditions are defined in a flexible manner.
 Optimization results are stored in a database.
-BlueM.Opt integrates a powerful function graph with the optimization results are displayed and can be analyzed.
+BlueM.Opt integrates a graphing feature for displaying the optimization progress and results. Optimization results can also be analyzed in detail.
+Where possible, BlueM.Opt utilizes multithreading in order to evaluate multiple parameter sets simultaneously.
 List of available methods (optimization algorithms):
 * '''[[PES]]''': Parametric Evolution Strategy
 * '''[[CES]]''': Combinatorial Evolution Strategy
 * '''[[HYBRID]]''': Combination of PES and CES
 * '''[[Hooke & Jeeves]]''': Hillclimbing Algorithm
-* '''[[MetaEvo]]''': multi-criterion, hybrid optimization algorithm
+* '''[[MetaEvo]]''': multicritera, hybrid optimization algorithm
 * '''[[DDS]]''': Dynamically Dimensioned Search
 another method is
 * '''[[SensiPlot]]''': Sensitivity analysis (no optimization)
+* '''[[Traveling Salesman]]'''
+List of currently implemented simulation models:
+* '''[[BlueM.Sim]]'''
+* '''[[SMUSI]]'''
+* '''[[S:CAN]]'''
+* '''[[SWMM]]'''
+* '''[[Test problems]]'''
 ==Downloads==