An Analysis of the Taguchi Method for Tuning a Memetic Algorithm with Reduced Computational Time Budget

Abstract

Abstract. Determining the best initial parameter values for an algorithm, called parameter tuning, is crucial to obtaining better algorithm performance; however, it is often a time-consuming task and needs to be performed under a restricted computational budget. In this study, the results from our previous work on using the Taguchi method to tune the parameters of a memetic algorithm for cross-domain search are further analysed and extended. Although the Taguchi method reduces the time spent finding a good parameter value combination by running a smaller size of experiments on the training instances from different domains as opposed to evaluating all combinations, the time budget is still larger than desired. This work investigates the degree to which it is possible to predict the same good parameter setting faster by using a reduced time budget. The results in this paper show that it was possible to predict good combinations of parameter settings with a much reduced time budget. The good final parameter values are predicted for three of the parameters, while for the fourth parameter there is no clear best value, so one of three similarly performing values is identified at each time instant.

Environmental and social implications of incorporating carpooling service on a customized bus system

“…With parameter tuning, the value of parameters is specified before executing a heuristic/meta-heuristic algorithm. If parameter tuning is used, the value of parameters is established in the initialization stage and remain unchanged during execution, for example Taguchi method (Roy, 1990;Gümüş et al, 2016;Sazvar et al, 2016), Response Surface Methodology (Montgomery, 2003;Myers et al, 2009), andIRACE (Dell'Amico et al, 2016;Samà et al, 2016).…”

Evolutionary Algorithm-Based Iterated Local Search Hyper-Heuristic for Combinatorial Optimization Problems

“…sequence-based selection hyper-heuristic (SSHH) and AdapHH emerged as the best two hyper-heuristics based on the experimental results. The new problem domains were used in two other works [26,27] that tuned a memetic algorithm, but only [27] compared the proposed hyper-heuristic with others that were tested on the domains, using 15 out of the entire set of 30 benchmarking instances.…”