Combining Environment-Driven Adaptation and Task-Driven Optimisation in Evolutionary Robotics
Figure 13
Effect of varying reward levels.
The plot shows histograms of green puck ratios across the population for different premium values for green pucks. Populations with a ratio of 0.5 are perfectly balanced, more extreme values indicate specialisation of the whole population on either type of puck. All plots report on the final 1,000 time steps of simulation. Adding a premium factor to green pucks causes a slight preference for collecting green pucks, but this effect is not very large.