Analytics and Machine Learning in Scheduling and Routing Optimization
Production scheduling and vehicle routing are two of the most studied fields in operations research. In the past four decades we have witnessed significant advances in both fields. However, the majority of existing research in both domains uses optimization based models and methodologies such as integer programming, dynamic programming and local search. In optimization, a problem is usually formulated into a mathematical model embedded with innate problem structures and characteristics. Such modeling and solution methods require the values of problem parameters to be available (i.e. in the form of either their deterministic values or their stochastic distributions) before the underlying mathematical models can be formulated and solved. However, real-life problems often involve a large amount of data which often contains a lot of uncertainty and changes over time. Optimization methods are often criticized for their inflexibility or ineffectiveness to deal with complex problems involving a large amount of data or a high degree of data uncertainty.
Analytic approaches, on the other hand. are entirely driven by data and often do not rely on rigid optimization models. Although such methods are more flexible than optimization methods, the resulting models and solutions have poor interpretability and may lack of insights that can be easily explained and understood by human users.
In the past several years, there has been growing research effort that attempts to bridge the gap between optimization and analytics, including methods that integrate optimization and machine learning.
This special issue aims to promote the use of this type of modeling and solution methods in production scheduling and vehicle routing. Specifically, we are seeking high quality scheduling and routing research papers that develop or apply integrated analytics and optimization methods that are not only flexible and robust under uncertainty, but can also generate models and solutions that are insightful and (relatively) easy to interpret. We are especially interested in papers that use one or more of the following modeling and solution methods: robust optimization, approximate dynamic programming, simulation optimization, stochastic programming, integer programming, and meta-heuristics, and their integration with data analytic tools such as optimal learning, machine learning, neural networks, and data mining. We are open to any interesting scheduling and routing applications including problems that arise in traditional areas such as production scheduling, vehicle routing, as well as applications from emerging areas such as supply chain scheduling, healthcare operations scheduling, routing with drones, ride sharing etc.
Authors wondering whether their research project is a fit for the special issue are encouraged to email a short description (no more than one page) of their project to the co-editors. We will provide feedback on whether the topic meets the goals of the special issue, although we will not evaluate the quality of the research based on the description because this will be left to the review process. There is no requirement to submit a description before submitting a paper.
Timeline and Process:
Deadline for submission: September 1, 2020
Second-round submission (for the papers invited to revise): January 1, 2021
Final decisions (subject to minor revisions): April 1, 2021
Guest Editors:
Ruibin Bai,
University of Nottingham,
Ningbo, Zhejiang Province, China
Ruibin.BAI@nottingham.edu.cn
Zhi-Long Chen,
University of Maryland,
College Park, MD 20742, USA
Graham Kendall,
University of Nottingham,
UK & Malaysia
调度和路径优化中的分析和机器学习
生产调度和车辆路线是运营研究中研究最多的两个领域。在过去四十年中，我们目睹了这两个领域的重大进展。但是，两个领域现有的大多数研究都使用基于优化的模型和方法，如整数规划，动态规划和局部搜索。在优化中，问题通常用有固有问题结构和特征的数学模型表示。在能被建立和求解之前，这种数学建模方法需要提供问题参数的值（即，以确定值或随机分布的形式）。然而，现实问题通常涉及大量数据，这些数据通常包含许多不确定性并且随时间而变化。优化方法经常因为在处理涉及大量数据或数据高度不确定性的复杂问题时的不灵活性或无效性而受到批评。
另一方面，分析方法完全由数据驱动，通常不依赖于严格的优化模型。虽然这些方法比优化方法更灵活，但是得到的模型和解决方案具有较差的可解释性，并且可能缺乏能够被用户轻松解释和理解的见解。
在过去几年中，越来越多的研究工作试图缩小优化和分析之间的差距，包括集成优化和机器学习的方法。
本特刊目的是促进在生产调度和车辆路径中使用此类建模和解决方法。具体而言，我们正在寻求高质量的调度和路径研究论文，这些论文开发或应用集成分析和优化方法，这些方法不仅在不确定性下具有灵活性和稳健性，而且还可以生成具有洞察力且（相对）易于理解的模型和解决方案。我们对使用以下一种或多种建模和解决方法的论文特别感兴趣：鲁棒性优化，近似动态规划，仿真优化，随机规划，整数规划和元启发式，以及它们与数据分析工具的集成，如优化学习，机器学习，神经网络和数据挖掘。我们对任何有趣的调度和路径应用程序持开放态度，包括传统领域中出现的问题，例如生产调度，车辆路径，以及来自新兴领域的应用，例如供应链调度，医疗保健运营调度，无人机路径，共享乘车等。
作者想知道他们的研究项目是否适合特刊，我们鼓励他们将他们项目的简短描述（不超过一页）通过电子邮件发送给共同编辑。我们将就该主题是否符合特刊的目标提供反馈，尽管我们不会根据描述评估研究的质量，因为这将留给审查过程。在提交论文之前无需提交描述。
时间表和流程：
- 提交截止日期：2020年9月1日
- 第二轮提交（邀请修改的论文）：2021年1月1日
- 最终决定（可做小幅修改）：2021年4月1日
特邀编辑：
Ruibin Bai,
University of Nottingham,
Ningbo, Zhejiang Province, China
Ruibin.BAI@nottingham.edu.cn
Zhi-Long Chen,
University of Maryland,
College Park, MD 20742, USA
Graham Kendall,
University of Nottingham,
UK & Malaysia