This paper considers how an online food delivery platform can improve last-mile delivery services’ performance using multi-source data. The delivery time is one critical but uncertain factor for online platforms that also regarded as the main challenges in order assignment and routing service. To tackle this challenge, we propose a data-driven optimization approach that combines machine learning techniques with capacitated vehicle routing optimization. Machine learning methods can provide more accurate predictions and have received increasing attention in the operations research field. However, different from the traditional predict-then-optimize paradigm, we use a new smart predict-then-optimize framework, whose prediction objective is constructed by decision error instead of prediction error when implementing machine learning. Using this type of prediction, we can obtain a more accurate decision in the following optimization step. Efficient mini-batching gradient and heuristic algorithms are designed to solve the joint order assignment and routing problem of last-mile delivery service. Besides, this paper considers the mutual effect between routing decision and delivery time, and provides the corresponding solution algorithm. In addition, this paper conducts a computational study and finds that the proposed method’s performance has an approximate 5% improvement compared with other methods.
The suddenness, complexity, and devastation of major geological disasters make it necessary to improve the efficiency of disaster rescue. At present, the dispatch of geological disaster rescue equipment represented by landslide and debris flow has some shortcomings, such as low efficiency, low matching degree and low utilization rate. Rescue equipment is an important guarantee for the rescue work. Predicting the demand for rescue equipment is conducive to optimizing the equipment scheduling decision, so as to improve the rescue ability of geological disasters. Based on the idea of ''disaster scenario-rescue mission-rescue equipment-demand forecast'', this study analyzes 87 representative geological disaster rescue cases in China from 2004 to 2019, and summarizes the matching relationship of ''scenario-task-equipment''. On this basis, The scenarios affecting the demand for various kinds of equipment were introduced as influencing factors, and the prediction models of rescue equipment demand based on different ''scenario-tasks'' were constructed; Finally, taking the rescue operation of ''7-23'' massive landslide in Shuicheng, Guizhou province in 2019 as an example, the demand prediction of rescue equipment was realized by adjusting parameter values, and the results were compared with the actual data of the case to verify the effectiveness and feasibility of the model.
Throughout the digitization of the petrochemical industry, the Beidou technology-based disaster monitoring, evaluation, and early warning network system has supported emergency decision making for oil and gas accidents. Many problems arise throughout the emergency decision-making process during oil–gas accidents, such as the limited time for decision making, high complexity, and inadequate emergency plans. Targeting these issues, we propose the construction of a case library using a Bayesian network. This way, when a new accident occurs, its similarity and deviation indexes could be matched against those of historical cases registered in the database. As such, the candidate cases are adjusted using a case combination and pruning method, yielding the final qualified case model. A case verification of the “11.22” Sinopec Oil pipeline leak and explosion in Qingdao reveals that the proposed method only requires an oil and gas accident database to be built in advance, eliminating the need for sampling data to make decisions, and reducing the search space. Using the proposed case-based reasoning, historical data and experience regarding oil and gas emergency decisions can be activated and reused, which would greatly improve the modeling efficiency of the Bayesian network.
The sudden complexity and destruction of major geological disasters determine the necessity of improving disaster relief efficiency. However, due to the relative lack of geological disaster relief data, the research on the relief efficiency evaluation of major geological disasters is insufficient. Based on the data of 18 rescue operations of major geological disasters in China from 2015 to 2019, this paper measure the disaster relief efficiency and analyze the influencing factors by using the super-efficiency SBM model and Tobit regression model with non-expected output. The research suggests several results: the average efficiency of rescue operations for 18 major geological disasters was 0.671, and that of disasters below the average efficiency was 67%; The average efficiency of several geological disaster relief operations in 2019 was 0.73, which explaining that after the establishment of the ministry of emergency management, the efficiency of geological disaster relief in China has been improved obviously; However, the long rescue period, redundant rescue personnel and unreasonable allocation of rescue equipment are the reasons why the rescue efficiency is relatively low for heavy and severe geological disasters. The research also found that the number of disaster victims and people trapped, the per capita GDP, the total population of the county, the county area and the highway density have a significant positive relationship with the rescue efficiency, whereas the disaster time, population density and highway mileage are negatively correlated with the rescue efficiency. In view of the above problems, policy suggestions are put forward to further improve the relief efficiency of China's major geological disasters.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.