Perbandingan Kinerja Algoritma Modified MSA (Method Of Successive Average) untuk Pembebanan Jaringan Jalan

Abstract

One of the crucial stages in the four-step model is traffic assignment model which is based on the user equilibrium (UE) principle. Mathematically, the UE assignment can be solved using the convex combination method in which one critical step is the determination of the direction-finding. The method of successive averages (MSA) is the basic method for determining the direction-finding. However, MSA faces long convergence times for large networks and large error for assignments that requires higher accuracy. This paper investigates two modifications of the standard MSA to address these issues, namely the method of successive weighted averages (MSWA) and the self-regulated averaging (SRA) method. These methods were then tested on Sioux Fall network and the road network of Bandung city as a case study respectively. The test results show that the MSWA method can improve the accuracy of the traffic assignment results, while the SRA method can reduce computation time.