Setting section

From SolutionSpaceTool

This page descripes the different settings and parameters in the Planer Tab.

Contents

General settings

General settings are the same for all algorithms.

z-D

This parameter controls the third dimension for only 2D benchmarks.

  • auto: The depth will be calculated upon the blocks in the benchmark.
  • manual: All blocks will have the given depth.

cost criteria

These are the cost criteria which can be calculated for a specific floorplan. One of these must be selected.

weight factors

These factors correspond with the cost criteria above. The calculated sum of the criteria is normalized to 1. With these factors the influence of one criteria can be adjusted. The sum of the weight factors must be 1 otherwise there occurs an error.

norm values

To calculate the sum of the cost criteria they must be normalized. The normalization is mostly done to the mean value of the criteria found by a Monte Carlo search. This values are saved in a database. If the database does not exist or the norm value was not yet searched, the program will automaticly switch to the estimate option and runs a short search for the mean value.

  • database: Use norm value from database.
  • estimate: Runs a short search for the mean value.

samples for norm value

This parameter sets the limit for the search of the mean value.

settings of Evolutionary Algorithm

parents

This parameter sets the number of parent individuals for each generation.

children

This parameter sets the number of children created by a recombination of two parent individuals.

generations

This parameter sets the number of generations created.

settings of Great Deluge Algorithm

vapor samples

This parameter sets the samples of searched solution after a vaporization before end the optimization

vapor value

This parameter sets the decreasing value of the critical value after a better solution was found.

settings of Monte Carlo Algorithm

samples

This parameter sets the number of random solutions.

settings of Simulated Annealing Algorithm

correct factor

This is a value for the difference of the sum between two solutions. At the temperatur Tend solutions with this difference will be accepted with a propability of 10%. To find the suitable option for this parameter you must run several different settings.

Tsteps

This parameter sets the steps between Tstart and Tend.

Tsamples

This parameter sets the solutions created each step.

Tstart

This parameter sets the start temperature.

Tend

This parameter sets the end temperature.

Personal tools