Monte Carlo Simulation for Trip Rate Estimation Using Genetic Algorithm-Based Stratified Sampling: Application in Low-Population Areas

Transport planning requires an accurate estimate of household trip rates to evaluate transportation needs. However, it has been believed that high accuracy is associated with large sample sizes, which hinders low-population local government units from conducting household travel surveys (HTS). For this reason, there is a need for a comprehensive understanding of how the trade-off between accuracy and sample size reduction behaves, using a stratified sampling approach to improve demographic representation, thereby reducing sample requirements in barangay-level communities. This study developed a genetic algorithm (GA)-based stratified sampling approach to optimize HTS sampling rates in both urban and rural low-population communities. Monte Carlo simulations with one million iterations evaluated the performance of GA-optimized sampling rates against random sampling. The stratification consistently achieved lower values in terms of Mean Absolute Percentage Error (MAPE), Standard Error, and Coefficient of Variation at optimized sampling rates, 7.82% for urban barangay and 9.19% for rural barangay, compared to random sampling at 8% and 10% sampling rates, respectively. Probability accuracy analysis further confirmed GA’s reliability in estimating trip rates at ≤10% MAPE accuracy, indicating high reliability above 60% empirical probability. Ultimately, this study demonstrates a risk-based approach to sample size selection in transportation studies.