Comparison of resource distribution metrics on multi-resource projects

Abstract

Abstract:Resource leveling is implemented to reduce the fluctuations of the daily resource usage and the peak resourceusage throughout the construction period. Measurement of peak resource demand is straightforward, whilefluctuations of resource distribution are difficult to evaluate. Therefore, many resource distribution metricsare proposed in order to measure the fluctuations of the resource distribution. In this study, minimummoment, absolute difference, sum of the squares, and minimum deviation resource distribution metrics arecompared by implementing resource distribution on four case study problems obtained from the previousstudies. The case study problems are optimized by exhaustive enumeration in order to ensure that the globaloptimum is obtained for each resource distribution metric. Recently the comparison was conducted for smallprojects. In this study, larger projects are evaluated to examine the effects of the project size and the searchdomain. Optimized resource distribution metrics are compared in terms of peak resource demand and theoverall resource distribution histogram. Similarity to the standard 'S' curve is expected for good fit resourcedistribution. The metrics are also compared in terms of easiness of implementation and computationaldemand. Results of the analyses revealed that minimum moment metric provided lower peak resourcedemand among the other resource distribution metrics. In addition to this, computational demand andeasiness of implementation of the minimum moment metric is the most suitable. The results of this researchcan be beneficial for the project managers through the selection of appropriate resource distribution metric.