Create a RandomSearchAlgorithm

[pfleck]

Having a random search algorithm can sometimes be usable, for example

• testing if a problem is simple enought that random sampling finds good solutions or
• generate some random sample data for learning a surrogate model for a more complex problem.

[pfleck]

r13580 Implemented Random Search Algorithm.

[pfleck]

r13581 Fixed output path of assembly.

[mkommend]

Reviewed r13580 and r13581.

Please rename the parameters to be more consistent with other algorithms (e.g., maximum iterations, maximum evaluated solutions, ...).

[abeham]

I would propose to call it MaximumEvaluations

[pfleck]

r13677: Renamed some parameters to be more consistent with other algorithms.

• Number of Samples -> Evaluated Solutions
• Batch Size -> Sample Size
• Number of Batches -> Iterations

@abeham I saw your suggestion just after the commit. MaximumEvaluations would also be fine. However, I choose MaximumEvaluatedSolutions because OSGA and CMAES also calls it that way.

[mkommend]

r13679: Corrected grammar mistake in RandomSearch.

[pfleck]

r13680

• Renamed SampleSize to BatchSize.
• Fixed name of iterations in the results in operator graph.
• Removed automatic recalculating of evaluated solutions or batch size when changing the maximum iterations.

[ascheibe]

r13863 added linux prebuild event

[mkommend]

r13952: Merged r13580, r13581, r13677, r13679, r13680, r13863 into stable.