Comparison of SOM-based optimization and particle swarm optimization for minimizing the construction time of a secant pile wall

Abstract

Optimization of construction time helps practitioners save time and money. In this study we compare two optimization methods — self organizing map based optimization (SOMO) and particle swarm optimization (PSO) — for minimizing the construction time of a secant pile wall. The comparison is based on data from 207 primary and secondary bored piles for a secant pile wall. The detailed construction time is measured in minutes and broken down into 16 work activities for each unit of the wall which is then used to yield the optimal construction sequence and time. Both optimization methods can be computed quickly but, SOMO yields a more efficient construction sequence with a shorter construction time. The comparison leads to several findings. Particles with randomly selected velocities and positions may lead to convergence at a local optimization. The PSO comparison mechanism for the yielding of optimization limits the process to the winner's neighbors, which may also converge to a local optimization. The case study shows the superiority of SOMO and provides answers to the dilemmas using the smallest hyper-plane, the entire comparison mechanism to yield the optimization, and the mechanism for weight adjustment between the winner neuron and its neighbors.