Simulation and KPI collection for assessment of the proposed on-demand dynamic routing based bus service by LTA
Using SimPy, created a simulation environment for the waiting times at "a typical bus stop" under the proposed SLA guidelines (Poisson arrival rates, exponential waiting times, discrete uniform random patience and reneges) Using Google ORTools API for Python, created a closed graph and minimum cost flow problem for closed bus route 255, using random poisson customers at each stop given a simulation based on the demand data from LTA DataMall. Each simulation then consisted of: setting up demand (source) points at random stops and departure (sink) points at random stops along the network. Then, using Min. Cost Flow solve the problem and record number of total buses required, total travelling time and operation cost against total distance traveled in network. Finally collected the results from the simulation studies and performed a comparative analysis of the present situation with the simulated results (after implementation)