Scholarlink

<p>This model is based on an interaction between a Samurai and a group of Assassins attempting to assassinate the Samurai, while he tries to cross a defined distance by defeating them and reaching the safe zone is simulated using an Agent-Based Model (ABM). This simulation determines whether the Assassins succeed in their mission or the Samurai reaches the safe zone can provide knowledge into group dynamics, decision-making under pressure, and leadership influence of the assassins. A comparison with existing Pursuit-Evasion models such as Predator–Prey model and Cops-Robbers model was made to find Success Probability and Efficiency in Decision-Making. The results shows that Success probability is higher compared to both the Predator–Prey model and Cops-Robbers model in equal or lesser decision steps. Decision tree was used to identify the stages of Samurai escaping or Assassins Succeeding, and 100 stages are recorded for predicting outcomes using Machine Learning. Random Forest was used for feature importance ranking and Gini Index score. Then, the Support Vector Machine (SVM) was used to train and make predictions based on those features. The actual and predicted outcomes are evaluated using confusion matrix, the results showed the accuracy score of 93.42, Kappa score of 0.8523, Matthews Correlation Coefficient as 0.7254, R-Squared score of 0.9377, MSE score of 0.0018 and finally p-value of 0.02455 (P &lt; 0.05) proves that the model statistically significant. For real-time testing Generative Adversarial Network (GAN) was used to create a dataset based on the Samurai-Assassins model to find Success statistics of an employee with Flex plot visualization using JASP.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Samurai and assassins strategy using agent based modelling and simulation for finding success probability and efficiency in complex scenarios

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月11日

查看全文
  • Vincent Gnanaraj T.,
  • Chenni Kumaran J.

摘要

This model is based on an interaction between a Samurai and a group of Assassins attempting to assassinate the Samurai, while he tries to cross a defined distance by defeating them and reaching the safe zone is simulated using an Agent-Based Model (ABM). This simulation determines whether the Assassins succeed in their mission or the Samurai reaches the safe zone can provide knowledge into group dynamics, decision-making under pressure, and leadership influence of the assassins. A comparison with existing Pursuit-Evasion models such as Predator–Prey model and Cops-Robbers model was made to find Success Probability and Efficiency in Decision-Making. The results shows that Success probability is higher compared to both the Predator–Prey model and Cops-Robbers model in equal or lesser decision steps. Decision tree was used to identify the stages of Samurai escaping or Assassins Succeeding, and 100 stages are recorded for predicting outcomes using Machine Learning. Random Forest was used for feature importance ranking and Gini Index score. Then, the Support Vector Machine (SVM) was used to train and make predictions based on those features. The actual and predicted outcomes are evaluated using confusion matrix, the results showed the accuracy score of 93.42, Kappa score of 0.8523, Matthews Correlation Coefficient as 0.7254, R-Squared score of 0.9377, MSE score of 0.0018 and finally p-value of 0.02455 (P < 0.05) proves that the model statistically significant. For real-time testing Generative Adversarial Network (GAN) was used to create a dataset based on the Samurai-Assassins model to find Success statistics of an employee with Flex plot visualization using JASP.