Multitask Scheduling in Consideration of Fuzzy Uncertainty of Multiple Criteria in Service-Oriented Manufacturing

Abstract

Tasks in the field of service-oriented manufacturing (SOM) such as cloud manufacturing have the characteristics of complexity, heterogeneity, uncertainty, and geographically distribution, which make scheduling them nontrivial and challenging, especially in the fuzzy environment. A fuzzy multicriteria modeling is of importance for the problem of fuzzy scheduling in SOM. In this article, four comprehensive models are proposed, which are different in the uncertain degree of considered performance criteria and/or defuzzification timepoints of fuzzy values. For each model, all weighted criteria are aggregated by using an exponential benefit function. For solving the models, three scheduling algorithms, namely one-level fuzzy ant colony optimization (OFACO), two-level single optimization fuzzy ACO (TSFACO), and two-level double optimization fuzzy ACO (TDFACO), are proposed. OFACO takes the view of the whole set of tasks on the SOM platform only whereas TSFACO and TDFACO consider both the view of the whole set of tasks and the view of individual task. The performance and effectiveness of the proposed fuzzy models and scheduling schemes are compared, respectively, using test datasets with varying sizes. The test results show that the first model with fuzzy objective is better for type-1 fuzzy uncertainty whereas the second model with defuzzified objective is better for type-2 fuzzy uncertainty and TDFACO outperforms the other two scheduling schemes in terms of the proposed integrated fuzzy multicriteria performance both from the individual task perspective and the whole task perspective.