Fuzzy back-propagation network approach for estimating the simulation workload. Issue 6 (August 2016)

• Record Type: Journal Article
• Authors: Toly Chen, Tin-Chih
• Is Part Of: Neural computing & applications. Volume 27:Issue 6(2016)
• Page Start: 1707
• Page End: 1715
• Publication Date: 2016-08
• Subjects: Simulation -- Workload estimation -- Fuzzy -- Back-propagation network; Neural networks (Computer science) -- Periodicals; Neural circuitry -- Periodicals; Artificial intelligence -- Periodicals; Neural Networks (Computer) -- Periodicals; Réseaux neuronaux (Informatique) -- Périodiques; Réseaux nerveux -- Périodiques; Intelligence artificielle -- Périodiques; 006.32
• DOI: 10.1007/s00521-015-1967-z
• Languages: English
• ISSNs: 0941-0643

Abstract

Abstract Estimating the time required for simulating a factory online is a crucial topic in manufacturing. However, this topic has rarely been discussed. A fuzzy back-propagation network (FBPN) approach for estimating the workload of a simulation task according to the required simulation time is proposed in this paper. In the proposed FBPN approach, tasks for which simulation times cannot be effectively estimated are considered as outliers and excluded, and therefore, more appropriate lower and upper bounds on the simulation time can be set. Thus, the ranges of all simulation times can be efficiently narrowed; however, this decrease is achieved at the expense of a slight decrease in the hit rate, which is still tolerable. A real case containing data of 90 simulation tasks was used to validate the proposed methodology. In addition, two existing FBPN methods, the adaptive-network-based fuzzy inference system method and the Chen (Comput Ind Eng 66:834–848, [2013]) method, were applied to these tasks for comparison. The experimental results showed that the proposed methodology was superior to the two existing FBPN methods in estimating precision and accuracy.