應用免疫球蛋白為基礎的演算法求解平行機排程

Applying Immunoglobulin-Based Algorithm for Parallel Machines Scheduling with Preference Machines

10.6270/JIBM.2013.4(2)141

廖麗滿(Li-Man Liao)；古天維(Tien-Wei Ku)

平行機 ； 免疫球蛋白為基礎的演算法 ； 抗體 ； 親和力 ； parallel machine ； immunoglobulin-based AIS algorithm ； antibody ； affinity

創新與經營管理學刊

4卷2期（2013 / 12 / 01）

141 - 161

繁體中文

電蝕鋁箔為鋁質電解電容器的主要元件，具有三種不同電壓設備的平行機，用來生產三種等級的產品，即高、中、低電壓之鋁箔。第一與第二種等級的設備都能用來生產第一（高電壓）與第二（中電壓）種等級的產品，若工作在不適合等級的機器處理，需要較長的處理時間。本研究發展免疫球蛋白為基礎的演算法（immunoglobulin-based AIS algorithm，IAIS），應用在鋁箔製造過程第一與第二等級設備之排程，初始抗體結合EDD法則，運用免疫球蛋白為基礎的演算法的多種突變機制，包含偏差銅板法、反轉與插入突變機制，求得目標函數為總延遲時間最小化。實驗結果顯示，求解績效優於彭郅絜（2011）提出的人工免疫演算法，績效改善率為11.64%。

The electro-etching aluminum foils are the major component in the aluminum electrolytic capacitor and are classified into different grade products. The electro-etching is the critical process in the manufacturing processes of aluminum foil, which usually has many parallel machines with three distinct voltage equipments used to produce three grades' products such as high voltage, medium voltage, or low voltage aluminum foil. In industrial practice, the first and second grade ones can used to produce the first (high voltage) and second (medium voltage) grade products alternatively. However, when jobs are not processed on their preferred machines, longer processing time are required. In this study, an immunoglobulin-based AIS (IAIS) algorithm is developed for the scheduling of the first and second grade equipments in the manufacturing process of aluminum foil. The objective function is to minimize total tardiness. The IAIS has a systematic immune mechanism which applies machine-based encoding method and integrates earliest due date (EDD) rule to find the initial antibodies. The coin toss mutation, insert mutation and swap mutation techniques are used to constructed antibody proliferation mechanism to obtain superior antibodies. To avoid falling into local optimal solutions, IAIS applies receptor editing mechanism to acquire the near-optimal scheduling. The computational results show that the performance of IAIS is significant superiority than Peng (2011)'s AIA. The improvement rate is 11.64%.