<![CDATA[The increasing demand for reliable lithium-ion battery in various applications is focused on the need for accurate State of Health (SOH) predictions to prevent performance degradation and potential safety risks. Therefore, this research aimed to improve the accuracy of SOH prediction by integrating Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) with Support Vector Machine (SVM) to overcome the overfitting problem in traditional machine learning models. The dataset used consisted of data from 1000 cycles of lithium-ion battery, collected under laboratory conditions. Data from lithium-ion battery cycles were analyzed using optimized PSO-SVM and ACO-SVM models. These models were evaluated using Mean Square Error (MSE) and Root Mean Square Error (RMSE) metrics, showing significant improvements in prediction accuracy and model generalization. The results showed that although both optimized models were superior to the baseline SVM, PSO-SVM had higher generalization performance during testing. The higher performance was due to the effective balance between exploring the search space and exploiting optimal solutions, making it more suitable for real-world applications. In comparison, ACO-SVM showed superior performance in training data accuracy but was more prone to overfitting, suggesting the potential for scenarios prioritizing high training accuracy. These results could be applied to extend the lifespan of lithium-ion battery, contributing to enhanced reliability and cost-effectiveness in applications.]]>
