基於代理模型之貫序實驗參數最佳化方法於染整業智慧製造應用案例

Surrogate-Based Sequential Parameter Optimization in Textile Dyeing Industry

莊雅筑(Ya-Chu Chuang) ； 林順傑(Shun-Chieh Lin) ； 陳芝安(Chih-An Chen) ； 董泯言(Min-Yan Dong) ； 陳秉洋(Ping-Yang Chen)

《中國統計學報》 59卷3期 (2021/09) Pp. 129-144

摘要

本研究提出在大量客製生產的製程下，如何利用統計模型快速且精確地確定製造配方。客製化產品的訂單需求多變且少量多樣，為了滿足客戶訂定的目標，在收到新訂單時，製造商須花費大量成本進行實驗來確定相對應的製造配方。因此，如何有效降低實驗成本是本研究之主要挑戰。我們採取一種基於多維代理模型的建模技術，可針對多變化的新訂單，尋找最佳製造配方。實務應用上，我們首先以布品染色製程為背景，介紹其中含三維目標顏色之染料配方實驗。接著，本研究基於該染料配方實驗設計模擬流程，以展示所提出模型之效能。結果顯示，我們提出的代理模型技術，對於各種不同的目標顏色，皆能準確預測出相對應的染料配方。

關鍵字

全域最佳化演算法 ； 高斯過程模型 ； 多維高斯過程模型 ； 智慧製造 ； 紡織染整製程

並列摘要

Our focus is on the manufacturing corporation that produces for a variety of customized products. Since the new incoming order frequently comes with different requirements, to satisfy the customers' targets, such corporation costs largely in the experiment to identify the proper manufacturing recipe for the product. Therefore, how to efficiently lower the cost of the experiments is the key challenge. In this work, we present a surrogate-based multivariate modeling technique that optimizes the manufacturing recipe for the new targets. To illustrate the idea, we first introduce the background of the color recipe experiment in the textile dyeing process where the three-dimensional target is involved. Then, we show the performance of our proposed technique by the simulation that mimics the experiment on the color recipe identification. The results show that the framework identifies good recipes efficiently for a wide range of the target color.

並列關鍵字

Efficient global optimization ； Gaussian process model ； Multivariate Gaussian process model ； Smart manufacturing ； Textile dyeing

參考文獻

Forrester, Alexander, Sobester, Andras and Keane, Andy (2008). Engineering design via surrogate modelling: a practical guide. John Wiley & Sons.

Jones, Donald R, Schonlau, Matthias and Welch, William J (1998). Efficient global optimization of expensive black-box functions. Journal of Global optimization, 13(4), pages 455-492.

Kennedy, James and Eberhart, Russell (1995). Particle swarm optimization. In Proceedings of IEEE International Conference on Neural Networks. IV, pages 1942-1948.

Nayak, Rajkishore and Padhye, Rajiv (2018). Artificial intelligence and its application in the apparel industry. In Automation in garment manufacturing, pages 109-138.

Park, J and Shore, J (2009). Evolution of right-first-time dyeing production. Coloration technology, 125(3), pages 133-140.