Research Article

Tactical Agent Personality

Figure 13

Generality test results for the SAP hierarchical learning framework: experiment sets 4, 5, and 6. The last three of six sets of experimental setups where each setup consists of Team B having a different fixed strategy. The exception is in the 6th set whereby Team B chooses a random strategy in each iteration. Each experiment set has the reward value plotted against the number of iterations. A single point on the graph represents a full game episode. Note that in order to maximize visibility, the scales are not uniform across all graphs.
107160.fig.0013a
(a)
107160.fig.0013b
(b)
107160.fig.0013c
(c)