Estimating Accurate Traffic Time by Smart Simulative Route Predictions

Dijkstra and other static shortest-path methods give optimal routes under fixed costs, but they miss evolving congestion, so ETAs drift. We propose a Smart Real-Time Estimation method that folds predictive models of future congestion into route evaluation.

We compare the method to standard Dijkstra on several networks and traffic levels. In medium traffic, anticipating congestion buildup improves ETA precision. In heavy congestion the approach reduces systematic error but variance rises with more complex interactions. Traffic signals affect both methods; ours adapts better when congestion patterns are predictable.

The result is a more dynamic routing estimate for navigation when conditions keep changing.