Updated with revised figures (script and output files)

Added a README to the analysis directory
Updated README with clearer information of the contents
Removed old output figures

README.md

@@ -26,14 +26,14 @@ srun julia archipelago.jl
 ## Code structure
 
 The wrapper script for a parallel run is `archipelago.jl`.
-Its single core equivalent is `singlecore.jl`.
+Its serial equivalent is `singlecore.jl`.
 
 Generated data for parallel runs can be found in the `data` directory.
-Single core data is in the `singlecore` directory.
+The serial approach data is in the `singlecore` directory.
 
 Utilities for data handling and plotting can be found in the `analysis` directory.
 
-All core additions to EvoLP.jl are in the `src/island.jl` file. The components are:
+All core additions to EvoLP.jl are in the `src/island.jl` file. These added components are:
 
 ### Deme selection blocks
 
@@ -99,7 +99,7 @@ A generational GA (`islandGA`) with:
 - Population size `popsize`: 30
 - Iterations: 100
 - Migration
-  - rate: `mu` (see below)
+  - rate: `mu` (see table below)
   - selection policy: `RandomDemeSelector` with `k=0.1*popsize`
   - replacement policy: `WorstDemeSelector` with `k=0.1*popsize`
 

analysis/README.md

@@ -0,0 +1,13 @@
+# Analysis scripts and tools
+
+This folder contains multiple tools and files for the analysis:
+
+- `analysis.jl`: a short script to join all CSV files in `data`. It generates DataFrame-ready CSVs for the parallel approach and their serial equivalent (with an `_s` at the end of the filename):
+    - `ackley.csv` and its serial equivalent: `ackley_s.csv`
+    - `eggholder`
+    - `michalewicz`
+    - `rana`
+    - `rosenbrock`
+- `plots.ipynb`: a Jupyter notebook for generating the figures used in the paper. It creates:
+    - `figure1.pdf`
+    - `figure2.pdf`