Game Development Reference
In-Depth Information
A famous algorithm used for creating decision trees from examples (and
that used in Black & White ) is ID3 developed by computer scientist
J.R. Quinlan in 1979.
Unity Hands On
Decision Trees
In this workshop we create a very simple decision tree for an RTS AI game
opponent that will decide on its strategy based on the map layout. The
decision being made is about whether or not to give priority to mining
for minerals and gas instead of building defenses or attacking the enemy.
Based on the amounts of nearby minerals and gases, the ease of entry
by which the enemy can access the base, and the overall distance to
the enemy, a table with all combinations of the values for each decision
variable (the attributes) is drawn up. Data for decision making are shown
in Table 5.5 .
In this case, data are fictitious; however, they could be gathered from
other human players or experience from previous games.
TABLE 5.5 Data for RTS AI build decisions
Rule
Number of
mineral
resources nearby
Number of gas
resources
Easy enemy
access to
base
Distance to
enemy
Give
priority
to mining
1
Low
High
No
Far
Ye s
4
Low
High
Ye s
Far
Ye s
7
Low
Low
No
Far
No
10
Low
Low
Ye s
Far
No
13
High
High
No
Far
Ye s
16
High
High
Ye s
Far
Ye s
19
High
Low
No
Far
Ye s
22
High
Low
Ye s
Far
Ye s
25
Low
High
No
Close
Ye s
28
Low
High
Ye s
Close
No
31
Low
Low
No
Close
No
34
Low
Low
Ye s
Close
No
37
High
High
No
Close
Ye s
40
High
High
Ye s
Close
Ye s
43
High
Low
No
Close
Ye s
46
High
Low
Ye s
Close
No
 
Search Nedrilad ::




Custom Search