See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/379603862

"IMPLEMENTATION OF BELLMAN-FORD ALGORITHM FOR DYNAMIC NPC

PATHFINDING IN HORROR GAMES" A Research Study Submitted to the
Department of Computer Studies at FEU Alabang In Partial Fu...

Experiment Findings · April 2024

DOI: 10.13140/RG.2.2.29115.22563

CITATIONS READS

0 318

9 authors, including:

Ma. Rovi Gwyneth Betonio Maxim Campos

FEU Alabang FEU Alabang
1 PUBLICATION 0 CITATIONS 1 PUBLICATION 0 CITATIONS

SEE PROFILE SEE PROFILE

Princess Orbillo
Far Eastern University
1 PUBLICATION 0 CITATIONS

SEE PROFILE

All content following this page was uploaded by Princess Orbillo on 06 April 2024.

The user has requested enhancement of the downloaded file.

“IMPLEMENTATION OF BELLMAN-FORD ALGORITHM

FOR DYNAMIC NPC PATHFINDING

IN HORROR GAMES”

A Research Study

Submitted to the Department of Computer Studies

at FEU Alabang

In Partial Fulfillment of the Requirements

In CCS0007 Algorithms

Betonio, Ma. Rovi Gwyneth

Campos, Maxim Luis M.

Enrique, Ranen Ezra Q.

Orbillo, Princess M.

Orteza, Luis Angelo M.

S.Y. 2023 - 2024

April 2024

1
 ABSTRACT

Pathfinding in video games is crucial for creating immersive and challenging gameplay

experiences, particularly in horror games where NPC behavior greatly contributes to atmosphere

and tension. The Bellman-Ford algorithm, known for its versatility and suitability, is adapted to

handle dynamic environments and address changing terrains and unpredictable player movements,

even with negative edges present in a weighted graph. This study investigates the implementation

of the Bellman-Ford algorithm in NPC pathfinding, considering dynamic factors. The

methodology involves integrating the algorithm into Roblox Studio and simulating the NPC's

behavior in response to obstacles, terrains, and the player's actions. Results indicate that the

Bellman-Ford algorithm demonstrated effective pathfinding, enabling NPCs to dynamically

navigate through the game world.

Keywords: Pathfinding, Non-player character (NPC), Bellman-Ford Algorithm and Roblox

Studio

2
 TABLE OF CONTENTS

TITLE PAGE……………………………………………………………………………………...1

ABSTRACT………………………………………………………………………………………2

INTRODUCTION………………………………………………………………………………...4

METHODOLOGY………………………………………………………………………………..7

RESULTS………………………………………………………………………………………12

RECOMMENDATION………………………………………………………………………...13

REFERENCES…………………………………………………………………………………14

3
 INTRODUCTION

Every successful video game features a few common elements, such as clear objectives,

restrictions, interactions, gameplay balancing mechanics, inertia, strategy, surprise elements,

characters, story, and music (Juegoadmin, 2023). Non-player characters, most commonly known

as NPCs, are one of them. According to Foudil et al. (2009) and Hill (2024), an NPC is a game

object not controlled by a player used to identify between characters the player controls and

characters the game controls completely. They are important in games as they assist in the

advancement of the main plot by interacting with players, offering helpful advice, and assigning

the main character to new objectives. NPCs also give the game vitality and add depth, complexity,

and realism, which greatly increases its excitement (Millington & Funge, 2009). In addition,

Montelli (2021) stated that NPCs hold importance in a game's narrative by portraying villains or

other important characters, enriching immersive gaming experiences with backstories, dialogue

options, and characteristics that resemble independent thinking.

In horror video games, non-player characters (NPCs) play a significant role in enhancing

both gameplay mechanics and narrative immersion, shaping the virtual world and enhancing the

player's journey (Millington & Funge, 2009). According to Yu (2022), NPCs contribute to

inducing feelings of fear, anxiety, or panic within players. By effectively integrating NPCs into

the game environment, developers aim to leave a lasting impact on players even after they have

concluded their gaming sessions. They serve as catalysts for eliciting fear and amplifying players'

immersive experiences through strategic integration within gameplay and narrative frameworks.

4
 The foundation of NPC behavior in games lies in the design of efficient pathfinding

algorithms, which are key components of the underlying Artificial Intelligence (AI) models

(Morina & Rafuna, 2023). As stated by Hunkeler et al. (2016), pathfinding is a plotting node by a

computer application to find the shortest route between two points from source to destination,

aiming to find the minimum total weight path in a graph between pairs of nodes (Cui & Shi, 2011).

Hunkeler et al. (2016) added that pathfinding plays a significant role in video game applications

as it is essential for developing realistic NPCs in static, dynamic, and real-time game environments.

These techniques used to determine the path in video games are called pathfinding algorithms,

which are used by the non-player character (NPC) to find a path between the starting point to the

desired point. Common pathfinding algorithms used in video games are Dijkstra and A*

algorithms. (Rafiq et al., 2020, p. 2). Pathfinding algorithms address the problem of solving for

the shortest and most efficient path from origin to destination and avoiding obstacles (Barnouti et

al., 2016).

When it comes to finding the shortest path in graphs with weighted edges, the Bellman-

Ford algorithm is another essential technique that programmers can use in pathfinding. The

Bellman-Ford algorithm is a pathfinding algorithm that computes the shortest path in a weighted

graph from a source vertex to all other vertices. Unlike other pathfinding algorithms, such as

Dijkstra’s, which can only work with non-negative weights, the Bellman-Ford is also designed to

handle negative weighted edges. It can recognize and manage negative weight cycles, in which the

sum of the weights along a cycle is negative. The main principle behind this algorithm is to

gradually update the distance estimations for each vertex while iteratively relaxing the graph’s

edges until the shortest paths are identified (Bhuyan, 2023). In video game development, the

5
Bellman-Ford algorithm assists NPCs in navigating through the game world efficiently. It helps

determine the optimal path for NPCs, considering obstacles, terrain, and other dynamic factors,

enhancing the realism and intelligence of in-game entities (Cloud Native Journey, 2023).

The developers aim to implement the Bellman-Ford algorithm for the dynamic pathfinding

of the non-player characters (NPCs) in horror games. This algorithm, known for its efficiency in

finding the shortest paths with negative edges present in weighted graphs, will be utilized to create

immersive experiences within the game environment. This will also assist the developers in

crafting suspenseful paths and unpredictable enemy movements, enhancing players' gaming

experience. By employing the Bellman-Ford algorithm, the NPCs will navigate through the

terrains and respond intelligently to dynamic conditions, adding depth and realism to player

interactions.

6
 METHODOLOGY

The developers, inspired by the narrative and gameplay elements highlighted in horror

video games, meticulously defined the eerie game environment within Roblox Studio, crafting its

layout, obstacles, and dynamic elements capable of perturbing NPC navigation. Drawing from the

rich descriptions provided in the introduction, such as inducing feelings of fear, anxiety, and panic,

the developers ensured that the game environment was intricately designed to evoke such

emotions.

Within Roblox Studio, they translated the complexities of this environment into a graph

representation, where each node symbolized a discrete location within the sinister world, while

edges represented connections delineating feasible pathways. These edges were assigned weights

based on factors like distance, terrain complexities, and other game dynamics, ensuring that NPCs

could navigate the treacherous landscape intelligently.

To implement the Bellman-Ford Algorithm, the developers first initialized distances from

the source NPC to all other nodes in the graph within Roblox Studio, setting them to infinity except

for the source node itself, designated as 0. They meticulously iterated through all edges,

systematically relaxing them to identify shorter paths to destination nodes. The process continued

for a set number of iterations, with careful attention paid to identifying and addressing negative

weight cycles within the Roblox environment.

Upon completion of iterations, distances to all nodes from the source NPC were updated

with the shortest paths, laying the foundation for optimized NPC navigation within the eerie game

7
environment. Leveraging Roblox Studio's capabilities, the developers orchestrated NPC

movement along these calculated shortest paths, ensuring NPCs adeptly navigated around

obstacles and dynamically adjusted their trajectories in response to environmental fluctuations.

To maintain immersion and suspense, the developers established a framework for dynamic

updates within Roblox Studio, continuously refining the graph representation and recalibrating

shortest paths as the game unfolded. This iterative process involved managing emergent obstacles

and shifts in NPC objectives to sustain the atmosphere of dread and uncertainty.

Furthermore, optimization efforts within Roblox Studio were undertaken to enhance the

efficiency of the pathfinding algorithm, including caching computed paths and employing spatial

partitioning techniques to minimize computational overhead. Rigorous testing within the Roblox

environment evaluated NPC pathfinding across diverse scenarios, ensuring alignment with the

horror-themed narrative and gameplay mechanics.

By adhering to this comprehensive methodology, the developers seamlessly integrated the

Bellman-Ford algorithm for dynamic NPC pathfinding, enhancing both gameplay mechanics and

narrative immersion to deliver a chilling gaming experience.

The provided code segments implement the Bellman-Ford algorithm for dynamic NPC

pathfinding in a Roblox game environment, specifically for killer and police NPCs. The

initialization phase defines spawn points for both types of NPCs and creates a graph representing

potential paths between these spawn points and various intersections within the game world. The

8
Bellman-Ford algorithm is then applied to calculate the shortest paths from the nearest spawn point

to the player's position.

To achieve dynamic pathfinding, functions such as `findNearestKillerSpawn` and

`findNearestPoliceSpawn` determine the closest spawn point based on the player's location.

Subsequently, the Bellman-Ford algorithm calculates distances and predecessors, constructing a

path from the player to the NPC's spawn point. Finally, the NPCs' movements are orchestrated

using the `SetPrimaryPartCFrame` method, adjusting direction and speed along the computed path

towards the player.

Overall, this integrated approach ensures efficient and adaptable NPC navigation, allowing

them to autonomously traverse the game environment while responding intelligently to changes in

the player's movements and environmental conditions.

9
Killer NPC Source Code:

10
Police NPC Source Code:

11
 RESULTS

The implementation of the Bellman-Ford algorithm for dynamic NPC pathfinding in horror

games yielded several notable results. Firstly, the algorithm demonstrated efficient pathfinding

capabilities, enabling NPCs to navigate through dynamic environments characterized by changing

terrains and obstacles. Its adaptability was evident as it successfully handled unpredictable player

movements, adjusting NPC paths in real-time to optimize navigation. Furthermore, Bellman-

Ford's ability to manage negative weighted edges ensured that NPCs could traverse challenging

terrain configurations without getting stuck, contributing to smoother gameplay experiences. The

algorithm's impact on realism and immersion was significant, with NPCs showcasing intelligent

responses to environmental changes, enhancing gameplay realism and immersing players deeper

into the horror-themed world. Despite the computational complexity inherent in the Bellman-Ford

algorithm, optimizations within Roblox Studio maintained satisfactory performance levels,

ensuring smooth gameplay experiences for players navigating through the game world alongside

dynamic NPCs.

12
 RECOMMENDATIONS

To enhance the implementation of the Bellman-Ford algorithm for dynamic NPC

pathfinding in horror games, several recommendations can be considered. Firstly, developers
should implement mechanisms for dynamic obstacle updating, allowing for the real-time
adjustment of obstacle positions and characteristics within the game environment. This capability
enables NPCs to react more realistically to changing conditions, enhancing the overall gameplay
experience. Additionally, continuous optimization of the pathfinding algorithm is crucial to reduce
computational overhead. Techniques such as caching frequently used paths and employing spatial
partitioning for efficient path calculations can significantly improve algorithm performance.

Providing user interface feedback is another essential recommendation. Visual or auditory

cues can be incorporated to notify players when NPCs change paths or encounter obstacles due to
dynamic pathfinding. This feedback enhances player awareness and immersion in the game world.
Scalability testing is also vital to ensure that the algorithm can handle larger and more complex
game environments while maintaining consistent performance and pathfinding accuracy.

Lastly, exploring ways to integrate player interactions into NPC pathfinding decisions can
create dynamic and engaging gameplay experiences. By incorporating these recommendations,
developers can further optimize the use of the Bellman-Ford algorithm, contributing to a more
immersive and engaging horror gaming experience overall.

13
 REFERENCES

Barnouti, N. H., Al-Dabbagh, S. S. M., & Naser, M. a. S. (2016b). Pathfinding in strategy games and

maze solving using a* search algorithm. Journal of Computer and Communications, 04(11), 15–25.

https://doi.org/10.4236/jcc.2016.411002

Bellman-Ford Algorithm: A pathfinding algorithm for weighted graphs. (2023, June 5). Cloud Native
Journey. https://cloudnativejourney.wordpress.com/2023/06/05/bellman-ford-algorithm-a-pathfinding-
algorithm-for-weighted-
graphs/#:~:text=The%20Bellman%2DFord%20algorithm%20helps,intelligence%20of%20in%2Dgame%
20entities.

Bhuyan, A. (2023, June 5). Bellman-Ford Algorithm: A pathfinding algorithm for weighted graphs.

Medium. https://aditya-sunjava.medium.com/bellman-ford-algorithm-a-pathfinding-algorithm-for-

weighted-graphs-647080392326

C. Foudil, D. Noureddine, C. Sanza, and Y. Duthen, “Path Finding and Collision Avoidance in Crowd

Simulation,” J. Comput. Inf. Technol., vol. 17, no. 3, p. 217, 2009.

Coyle, N. (2020b, November 19). The Psychology of horror games - Psychology and video games.

Psychology and Video Games. https://platinumparagon.info/psychology-of-horror-games/

Cui, X. and Shi, H. (2011) A*-Based Pathfinding in Modern Computer Games. International Journal of

Computer Science and Network Security, 11, 125-130.

Hill, S. (2024b, January 10). What is an NPC? Digital Trends.

https://www.digitaltrends.com/gaming/what-is-an-npc/

14
 I. Hunkeler, F. Schar, R. Dornberger, and T. Hanne, “FairGhosts - Ant colony controlled ghosts for Ms.

Pac-Man,” 2016 IEEE Congr. Evol. Comput. CEC 2016, pp. 4214–4220, 2016.

Juegoadmin. (2023b, September 22). Major game design elements. Juego Studio.

https://www.juegostudio.com/blog/game-design-elements

Millington, I., & Funge, J. (2009). Artificial Intelligence for Games. CRC Press.

Montelli, C. (2022b, August 23). What does “NPC” mean? Understanding non-player characters, an

important aspect of any video game. Business Insider. https://www.businessinsider.com/guides/tech/npc-

meaning

Morina, V., & Rafuna, R. (2023, October). A Comparative Analysis of Pathfinding Algorithms in NPC

Movement Systems for Computer Games. ResearchGate.

https://www.researchgate.net/publication/375895663_A_Comparative_Analysis_of_Pathfinding_Algorith

ms_in_NPC_Movement_Systems_for_Computer_Games

Rafiq, A., Kadir, T. a. A., & Ihsan, S. N. (2020b). Pathfinding Algorithms in game development. IOP

Conference Series: Materials Science and Engineering, 769(1), 012021. https://doi.org/10.1088/1757-

899x/769/1/012021

What is an NPC and Why It Matters for your Game Storytelling. (n.d.-c).

https://www.rosebud.ai/blog/what-is-an-npc-video-games-storytelling

15

View publication stats