Information Technology Reference
In-Depth Information
6
Evaluation
6.1
Evaluation Methodology
We developed a FPS game to implement our solution. Video demonstration of our algo-
rithm and playable version of our game are available at our project homepage [2]. We
evaluated the effectiveness of the evolutionary algorithm on producing good maps and
its compliance with the design goals.
6.2
Effectiveness of Evolution
We created 3 different maps using our solution. We run the evolution algorithm for 55
populations (approximately 10 seconds in a Intel Core i7 laptop) each time to generate
the map. The maximum average summed fitness is 4.0, as the fitness for each of the
4 features is normalised from 0.0 to 1.0. The average summed fitness of the processed
population against the number of populations processed so far is plotted in Figure 6.
The positive gradient shows that the evolutionary methods does improve the quality of
the map as more population are generated and mutated. It also shows that the fitness
stabilises after 26 populations (which takes about 5 seconds).
The fitness for each feature (with unnormalised values) of generated maps are shown
in Table 1. Based on the fitness function that we have defined, the maps are all con-
nected (with Connectivity of value 1.0), meaning that no regions are blocked from the
rest. The algorithm is also effective in constraining the number of collision points (with
Forced Collision Points of value 1.0, indicating that there are either 1 or 2 collision
points.). For both Flag Fairness and Overall Flag Fairness , we show the unnormalised
values. These values indicates the absolute difference in moving from the spawn points
to own team flags for Flag Fairness, and the difference in moving to all flags for Overall
Flag Fairness. A value of 1.0 means their distance are exactly one cell apart. Given our
result, the values are very low, with the highest being 1.03510, which is barely one cell
apart. To give a clearer perspective, one cell will take only approximately a second to
travel. Henceforth, we can conclude that the flags are placed in positions that are largely
fair for both teams.
Ta b l e 1 . Fitness values of generated
maps (after 26 populations)
Fig. 6. Change in fitness when evolving
 
Search WWH ::




Custom Search