algorithm parameter tuning is an often neglected step in the optimization process. This study shows that constrained multiobjective optimization can benefit significantly from tuning, in both the specialized (for an i...
详细信息
ISBN:
(纸本)9783031779404;9783031779411
algorithm parameter tuning is an often neglected step in the optimization process. This study shows that constrained multiobjective optimization can benefit significantly from tuning, in both the specialized (for an individual problem) and generalized (over a number of problems) parameter setting approaches. Numerical experiments conducted with three multiobjective optimization algorithms on 139 test problems from 13 benchmark suites quantify the algorithm performance improvement on individual problems. Additionally, regarding the generalized approach, alternative default parameter settings are identified. The study also identifies Bayesian optimization as an adequate method for tuning multiobjective evolutionary algorithms with constraint handling. Overall, it is concluded that, given sufficient computational resources to apply to a problem, parametertuning, using an approach such as Bayesian optimization, should be conducted. If computational resources do not allow such tuning, then the proposed default parameters are applicable.
This work presents an algorithm for tuning the parameters of stochastic search heuristics, the Robust parameter Searcher (RPS). RPS is based on the Nelder-Mead Simplex algorithm and on confidence-based comparison oper...
详细信息
ISBN:
(纸本)9783031232350;9783031232367
This work presents an algorithm for tuning the parameters of stochastic search heuristics, the Robust parameter Searcher (RPS). RPS is based on the Nelder-Mead Simplex algorithm and on confidence-based comparison operators. Whilst the latter algorithm is known for its robustness under noise in objective function evaluation, the confidence-based comparison endows the tuningalgorithm with additional resilience against the intrinsic stochasticity which exists in the evaluation of performance of stochastic search heuristics. The proposed methodology was used to tune a Differential Evolution strategy for optimizing real-valued functions, with a limited function evaluation budget. In the computational experiments, RPS performed significantly better than other well-known tuning strategies from the literature.
This study develops an enhanced ant colony optimization (E-ACO) meta-heuristic to accomplish the integrated process planning and scheduling (IPPS) problem in the job-shop environment. The IPPS problem is represented b...
详细信息
This study develops an enhanced ant colony optimization (E-ACO) meta-heuristic to accomplish the integrated process planning and scheduling (IPPS) problem in the job-shop environment. The IPPS problem is represented by AND/OR graphs to implement the search-based algorithm, which aims at obtaining effective and near-optimal solutions in terms of makespan, job flow time and computation time taken. In accordance with the characteristics of the IPPS problem, the mechanism of ACO algorithm has been enhanced with several modifications, including quantification of convergence level, introduction of node-based pheromone, earliest finishing time-based strategy of determining the heuristic desirability, and oriented elitist pheromone deposit strategy. Using test cases with comprehensive consideration of manufacturing flexibilities, experiments are conducted to evaluate the approach, and to study the effects of algorithmparameters, with a general guideline for ACO parametertuning for IPPS problems provided. The results show that with the specific modifications made on ACO algorithm, it is able to generate encouraging performance which outperforms many other meta-heuristics.
暂无评论