Project/Area Number |
25540172
|
Research Category |
Grant-in-Aid for Challenging Exploratory Research
|
Allocation Type | Multi-year Fund |
Research Field |
Entertainment and game informatics 1
|
Research Institution | Kwansei Gakuin University |
Principal Investigator |
|
Project Period (FY) |
2013-04-01 – 2016-03-31
|
Project Status |
Completed (Fiscal Year 2015)
|
Budget Amount *help |
¥3,770,000 (Direct Cost: ¥2,900,000、Indirect Cost: ¥870,000)
Fiscal Year 2014: ¥1,690,000 (Direct Cost: ¥1,300,000、Indirect Cost: ¥390,000)
Fiscal Year 2013: ¥2,080,000 (Direct Cost: ¥1,600,000、Indirect Cost: ¥480,000)
|
Keywords | ゲームプログラミング |
Outline of Final Research Achievements |
Designing the behavioral patterns of video game agents (NPC) is a crucial aspect in developing video games. While various systems that have aimed at automatically acquiring behavioral patterns have been proposed and some have successfully obtained stronger patterns than human players, those patterns have looked mechanical. This study proposes the autonomous acquisition of NPCs' behaviors, which emulate the behaviors of human players. Instead of implementing straightforward heuristics, the behaviors are acquired using techniques of reinforcement learning with Q-Learning and pathfinding through an A* algorithm, where biological constraints; "sensory error", "perceptual and motion delay", "physical fatigue", and "balancing between repetition and novelty" are imposed. The subjective assessments using ''Infinite Mario Bros'' indicated that agents with biological constraints could acquire more human-like behavioral patterns than humans and agents without them.
|