朱致成

Zhi-ChengZhu

碩士論文（2019）

適用於區塊鏈為基智能製造之協同式排程方法研究

Research on Collaborative Scheduling Method for Blockchain-based Smart Manufacturing

關鍵字 Keywords

區塊鏈、智能合約、智能製造、彈性零工式排程問題、決策樹

blockchain、smart contract、smart manufacturing、flexible job shop scheduling problem、decision tree

摘要

資訊技術的進步與發展，使得智能製造(Smart Manufacturing)得以實現。在智能製造中，所有的生產單元如設備、廠區等，皆能夠針對所遭遇的情境進行快速、正確的決策，以完成任務。排程是生產的重要活動之一，其任務是針對現有的資源，進行最佳的工作分派、製造資源分配、工作順序安排。由於傳統生產排程已無法滿足智能製造快速回應以及各級加工製造單元之自主性與協同合作性的需求，發展適用於智能製造之排程方法，已刻不容緩。而生產資料的保存、分享、正確與安全，在智能製造環境中至為重要，也是實務上待解決的問題。

針對適用於智能化生產之排程方法的需求，以及生產資料安全的問題，本研究分析智能製造之模式、參考區塊鏈技術，提出一以區塊鏈(Blockchain)為基適用於智能製造的協同式排程模式，透過智能機台之間相互溝通、協同來規劃排程。此模式包含協同式排程、協同式排程規則挖掘以及智能合約(Smart Contract)。

依此模式，首先運用決策樹(Decision Tree)分析方法，將歷史排程資料進行作業爭取規則與作業排序規則挖掘，並將規則寫入智能合約後部署上區塊鏈，以確保排程規則能夠安全無虞的被儲存與應用。在協同式排程中，機台針對生產需求，依自身之條件，選擇可以加工的作業。如遇多部機台有相同選擇時，則參考作業爭取規則，決定工作分派。當機台拿到各自的作業後，再參考作業排序規則，決定作業順序。本研究以實驗證明，所提之協同式排程方法能夠產出合理之排程，且能藉由不斷優化排程規則，達到累積經驗、優化排程的效果。

Abstract

In smart manufacturing, all production units, such as equipment and plant areas, can make quick and correct decisions for the situations encountered to complete the task. Scheduling is one of the important activities of production. Its task is to carry out the best job assignment, manufacturing resource allocation and work order arrangement for existing resources. Since traditional production schedules have been unable to meet the rapid response of smart manufacturing and the autonomy and synergy of processing and manufacturing units at all levels, it is imperative to develop a scheduling method suitable for smart manufacturing. The preservation, sharing, correctness and safety of production data are of paramount importance in the smart manufacturing environment, and they are also issues to be solved in practice.

In view of the requirements for scheduling methods for smart production and the safety of production data, this study analyzes the mode of smart manufacturing and reference blockchain technology, and proposes a collaborative scheduling model based on blockchain for smart manufacturing, through the smart machine to communicate and coordinate with each other to plan the schedule. This model includes collaborative scheduling, collaborative scheduling rule mining, and smart contracts.

Using the decision tree to mine historical scheduling data, find operation competition rules and operation sequencing rules, the machine can coordinate, make decisions, and plan scheduling through these two rules, and write the two rules into the smart contract, the deploy blockchain to ensure the safety of the rules.

This study proves that the proposed collaborative scheduling method can produce a reasonable scheduling, and can continuously accumulate scheduling rules to achieve cumulative experience and optimize scheduling.