Academia.eduAcademia.edu

Figure 10 – uploaded by Mohammad Zohaib

See full PDFdownloadDownload figure
3.6. Upper Confidence Bound for Trees and Artificial Neural Networks. Li et al. [36] developed a DDA technique using artificial neural networks (ANN) from data derived from the upper confidence bound for trees (UCT). The Pacman game was used as a test-bed for this study. Considering that UCT is a computing intelligence method, UCT perfor- mance significantly correlates with the duration of simulation [37]. Figure 8 illustrates DDA process from UCT-created data.

Figure 8 3.6. Upper Confidence Bound for Trees and Artificial Neural Networks. Li et al. [36] developed a DDA technique using artificial neural networks (ANN) from data derived from the upper confidence bound for trees (UCT). The Pacman game was used as a test-bed for this study. Considering that UCT is a computing intelligence method, UCT perfor- mance significantly correlates with the duration of simulation [37]. Figure 8 illustrates DDA process from UCT-created data.

Related Figures (14)

Ficure 1: Flow channel concept proposed by Csikszentmihalyi. TaBLE 1: DDA research studies since 2009.
Ficure 1: Flow channel concept proposed by Csikszentmihalyi. TaBLE 1: DDA research studies since 2009.
TABLE 2: List of DDA approaches.
TABLE 2: List of DDA approaches.
Ficure 3: Probabilistic graph showing the player’s progression model in a typical level-based game.
Ficure 3: Probabilistic graph showing the player’s progression model in a typical level-based game.
Ficure 4: Two-agent optimized levels of fun.
Ficure 4: Two-agent optimized levels of fun.
Ficure 6: The dynamic scripting process.
Ficure 6: The dynamic scripting process.
Ficure 7: A simple FPS game state transition illustration.
Ficure 7: A simple FPS game state transition illustration.
Figure 9 illustrates the DDA from ANN. The x-axis ranging within 0-40ms is the same as Figure 7. The y-axis represents win-rate of opponents (ghosts) controlled by ANN from data created by UCT with changed simulation time ranging within 20%-86%. The curve rises steeply in the range 0-100ms. After 100ms, the curve rises steadily and the win- rate peaks at 400ms.
Figure 9 illustrates the DDA from ANN. The x-axis ranging within 0-40ms is the same as Figure 7. The y-axis represents win-rate of opponents (ghosts) controlled by ANN from data created by UCT with changed simulation time ranging within 20%-86%. The curve rises steeply in the range 0-100ms. After 100ms, the curve rises steadily and the win- rate peaks at 400ms.
Ficure 10: Illustration of self-organizing algorithm.
Ficure 10: Illustration of self-organizing algorithm.
Related topics:
Computer ScienceHuman Computer Interaction
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved