Hits:

Indexed by:期刊论文

Date of Publication:2022-06-28

Journal:管理科学

Affiliation of Author(s):经济管理学院

Issue:2

Page Number:137-144

ISSN No.:1672-0334

Abstract:The great development of Internet of Things ( IoT) brings chances and challenges for production scheduling systems . Traditional scheduling optimization methods are usually based on human experience , mathematical models or the both .However, under IoT environment , the soaring multi-source and heterogeneous data , the apparent information , and the continuous and dy-namic existence of “humans, materials, facilities, production processes, products, etc.” in production scheduling systems disa-ble these traditional scheduling optimization methods . Based on literature review , the state-of-the-art of IoT, scheduling optimization methods and context-based modeling methods is summarized.For the field of IoT, specific scheduling problems should be considered further to apply the current general IoT the -ories to the practice .For the field of scheduling optimization , existing methods almost developed for structured or semi-structured problems which couldn′t resolve the complex and unstructured problems under IoT environment .For the field of context-based modeling, although existing methods pave the way for the development of context -aware systems, further study is still needed in the aspects of capturing and representing typical contexts from multi-source, heterogeneous, and massive data, and reasoning based on them to realize the modeling process to support the decision making .We conclude that the key scientific issue of IoT-based online intelligent scheduling is the context-based online modeling .The modeling process is “capturing context→represen-ting context→reasoning based on context”, which could realize the translation process of “data→information→model→schedu-ling policies”.Then, the future research goal is presented for dealing with the dynamic and continuous variations of scheduling objects under IoT environment , an online real-time and intelligent optimization method of scheduling should be invented to smooth the scheduling process and to provide scientific , efficient and practical decision support policies .Finally, the future research content is described in detail , which includes the following three aspects:context-based modeling methods , context-based and re-al-time model-solving methods , and context-based online decision support methods of scheduling .For the aspect of context-based modeling methods , the main research content includes the following:①the judgment of context series and the robustness analysis of them,②the representation of context series , and③the distributed modeling methods based on context series .As to context-based and real-time model-solving methods , the main research content includes the following:①online learning-feedback meth-ods based on context ,②distributed online and real-time model-solving algorithms , and③self-adaptive algorithms for distributed models.As to context-based online decision support methods of scheduling , the main research content includes the following:①cooperative and interactive decision-making methods , ②human-computer interactive methods based on context series , ③effec-tiveness and robustness analysis of scheduling decisions , and④the application research of scheduling decision support systems . This exploring work would pave the way for future research of the IoT-based decision support systems of scheduling .And the re-search results could have wide application in areas of production scheduling and logistics scheduling .

Note:新增回溯数据