Optimizing Incident Detection Thresholds Using the A* Algorithm: An Enhanced Approach for the California Algorithm

Abstract

This paper presents an improved version of the California Algorithm (CA), focusing on threshold selection criteria. The CA is a widely recognized incidence detection algorithm used as a benchmark for comparison with newly developed incident detection algorithms. This study proposes criteria for threshold selection in CA based on the A* algorithm, which aims to find optimal thresholds using a Performance Index (PI) as a cost function. Our proposed method reduces processing time by optimizing resource utilization and establishes a standard for threshold selection in CA for comparison and evaluation purposes. Experimental results from our proposed method demonstrate its effectiveness in reducing the complexity required to determine optimal thresholds. Optimization of the CA method using the A* algorithm results in a 98.68% reduction in the number of nodes searched compared to a Complete Search Tree (CST).