Pursuit-Evasion Game

A pursuit-evasion game is a branch of game theory that analyzes the dynamic strategies of two opposing parties—a pursuer and an evader—within a defined space. In automotive cybersecurity, it models the hacker as the 'pursuer' aiming to exploit vulnerabilities, and the vehicle's defense systems as the 'evader' aiming to detect and block attacks. This approach mathematically formulates the cybersecurity conflict, enabling analysis of optimal strategies under information asymmetry. It directly supports the Threat Analysis and Risk Assessment (TARA) methodology in ISO/SAE 21434, Clause 15, which requires systematic identification of attack paths and feasibility. The model provides a quantitative framework to simulate attacker decisions, allowing for more precise risk evaluation and control design than traditional static assessments.

In automotive cybersecurity risk management, the pursuit-evasion game model is applied through these steps: 1. **Threat Scenario Modeling**: Identify critical assets (e.g., ECUs, CAN bus) and attack vectors as per the ISO/SAE 21434 TARA process, structuring them as a game map. 2. **Define Strategies and Payoffs**: Quantify attacker strategies (e.g., port scanning) and defender strategies (e.g., dynamic resource allocation for monitoring). Establish a 'payoff matrix' based on the value of a successful attack to the hacker and the resulting loss for the enterprise. 3. **Nash Equilibrium Analysis**: Use algorithms to compute the Nash Equilibrium, revealing the most stable strategy profile for both players. This identifies the system's most probable weaknesses and the optimal allocation of defensive resources. A German OEM used this to reduce its response time to potential zero-day attacks by approximately 30%.

Taiwanese enterprises face three main challenges: 1. **Scarcity of Interdisciplinary Talent**: The model requires expertise in game theory, cybersecurity, and automotive engineering, a rare combination. Solution: Form cross-functional teams with internal experts and external consultants, initiating a 6-month talent development plan. 2. **Difficulty in Parameterization**: Accurately quantifying attacker costs, motivations, and defender resources is difficult without historical data. Solution: Start with qualitative inputs from frameworks like NIST SP 800-30 and internal TARA results, piloting the model on 1-2 high-risk scenarios. 3. **High Computational Cost**: Complex vehicle systems lead to large-scale models requiring significant computing power. Solution: Employ simplified or hierarchical models and leverage scalable cloud computing resources to avoid large upfront hardware investments.

Winners Consulting specializes in pursuit-evasion game for Taiwan enterprises, delivering compliant management systems within 90 days. Free consultation: https://winners.com.tw/contact