SoC Ideas AI Defense Strategies 2013 - Revision history

Crab: /* Additional Information */

2013-04-09T09:40:34Z

‎Additional Information

Crab: /* Description */

2013-03-26T20:49:40Z

‎Description

Crab: /* Description */

2013-03-26T20:44:48Z

‎Description

Crab: /* Description */

2013-03-26T20:33:56Z

‎Description

Crab: Created page with '{{Template:SoC2013Idea}} = Description =

AI: Implement a 'total defense' strategy

Page for the idea: SoC Ideas AI Defense Strategies 2013 AI in wesnoth's is good a…'

2013-03-26T20:33:13Z

Created page with '{{Template:SoC2013Idea}} = Description = <h2>AI: Implement a 'total defense' strategy</h2> Page for the idea: SoC Ideas AI Defense Strategies 2013 AI in wesnoth's is good a…'

New page

{{Template:SoC2013Idea}}
= Description =
<h2>AI: Implement a 'total defense' strategy</h2>

Page for the idea: [[SoC Ideas AI Defense Strategies 2013]]

AI in wesnoth's is good at killing things but is bad at defending. Therefore we are forced to make it attack at 2:1 odds. This is the single biggest limiter of the AI skill at the moment. To solve that, we need to teach the AI to defend.

{{#dpl:
|resultsheader=''There are %PAGES% submitted student proposals for this idea''
|oneresultheader=''There is 1 student proposal for this idea''
|suppresserrors=true
|noresultsheader=''There are no student proposals for this idea''
|category=Summer of Code 2013 Student Page&SoC Ideas AI Defense Strategies 2013
|notcategory=SoC 2013 Not Submitted To Google
|include=#Description
|nottitlematch=SoC2013_Template_of_Student_page
|mode=userformat
|format=,,<br/>See [[%PAGE%|%TITLE%]] for more information.<br/><br/>,
}}

=Additional Information=

Currently, the aim is to maximize (enemy losses - aggression*own_losses), where aggression is currently set to 0.4 by default. Teaching the AI to defend would allow us to lower the default aggression value. If we can bring the aggression value close to 0, the problem would be solved.

What would happen if we set aggression to 0 now? AI would only do the attacks which are 'good at killing things', but would not do the attacks, leaving only several units on the frontline. Moreover, the formula max(enemy_losses - aggression*own_losses) is broken, because it does not consider next_turn retaliation properly. A correct (but not calculable) formula would be max(enemy_losses_defending - aggression*own_losses_attacking - aggression*own_losses_next_turn + enemy_losses_attacking_next_turn + enemy_losses_defending_in_two_turns - aggression*own_losses_attacking_in_two_turns - … ). It's not calculable because we can't predict all the turns in advance with any degree of confidence. However, if we take the *next* turn into account better, then the AI would be able to play in styles other than 'all-out-attack' - because, if you take next turn into account, it makes sense to form defensive lines, block enemy units from reaching your vulnerable units, etc.

=Whom to ask about this=
Crab_ on irc.

← Older revision		Revision as of 20:49, 26 March 2013
Line 6:		Line 6:

	AI in wesnoth's is good at killing things but is bad at defending. Therefore we are forced to make it attack at 2:1 odds. This is the single biggest limiter of the AI skill at the moment. To solve that, we need to teach the AI to defend.		AI in wesnoth's is good at killing things but is bad at defending. Therefore we are forced to make it attack at 2:1 odds. This is the single biggest limiter of the AI skill at the moment. To solve that, we need to teach the AI to defend.
−

	Suggested skills: C++, knowledge of wesnoth's gameplay (can be learned during the project)		Suggested skills: C++, knowledge of wesnoth's gameplay (can be learned during the project)

← Older revision		Revision as of 20:44, 26 March 2013
Line 8:		Line 8:


−	~~Required~~ skills: C++, knowledge of wesnoth's gameplay (can be learned during the project)	+	Suggested skills: C++, knowledge of wesnoth's gameplay (can be learned during the project)

	{{#dpl:		{{#dpl:

← Older revision		Revision as of 20:33, 26 March 2013
Line 6:		Line 6:

	AI in wesnoth's is good at killing things but is bad at defending. Therefore we are forced to make it attack at 2:1 odds. This is the single biggest limiter of the AI skill at the moment. To solve that, we need to teach the AI to defend.		AI in wesnoth's is good at killing things but is bad at defending. Therefore we are forced to make it attack at 2:1 odds. This is the single biggest limiter of the AI skill at the moment. To solve that, we need to teach the AI to defend.
		+
		+
		+	Required skills: C++, knowledge of wesnoth's gameplay (can be learned during the project)

	{{#dpl:		{{#dpl:

@@ Line 27: / Line 27: @@
 What would happen if we set aggression to 0 now? AI would only do the attacks which are 'good at killing things', but would not do the attacks, leaving only several units on the frontline. Moreover, the formula  max(enemy_losses - aggression*own_losses) is broken, because it does not consider next_turn retaliation properly. A correct (but not calculable) formula would be   max(enemy_losses_defending - aggression*own_losses_attacking - aggression*own_losses_next_turn + enemy_losses_attacking_next_turn + enemy_losses_defending_in_two_turns - aggression*own_losses_attacking_in_two_turns - … ). It's not calculable because we can't predict all the turns in advance with any degree of confidence.  However, if we take the *next* turn into account better, then the AI would be able to play in styles other than 'all-out-attack' - because, if you take next turn into account, it makes sense to form defensive lines, block enemy units from reaching your vulnerable units, etc.
 =Whom to ask about this=
 Crab_ on irc.