讲座时间：2019年6月28日（星期五）上午10:00-11:00

讲座地点：主教615会议室

讲座一:

演讲人👩🏻‍🦳：孙睿卿（凯发平台2018级硕博连读生，技术经济及管理专业）

题目：Incentive alignment,collaborative carbon reduction in supply chains and performance

摘要：The purpose of this study isto explore the antecedents and consequences of incentive alignment inperforming collaborative carbon reduction in supply chains (CCR-SC). To fulfillthe research objective, an empirical study was carried out in China, anemerging economy that has in recent years started to undergo a transitiontowards a low-carbon economy. The statistical technique of structural equationmodelling (SEM) was used to analyze the questionnaire data collected from 120Chinese companies. The research results reveal that incentive alignment has apositive impact on CCR-SC, which is instrumental in the improvement of bothfinancial and environmental performance.

讲座二：

演讲人：余昊洋（凯发平台2018级学术型硕士，技术经济及管理专业）

题目𓀉：A clustering-based sales forecast method for big promotion days inO2O on-demand retailing

摘要：The O2O on-demand business is a fast-growing business modeldesigned to integrate traditional offline retailing and e-commerce. The salesforecasting for this business pattern is very important, especially in bigpromotion days, as it will directly influence the capacity plan and thereplenishment plan of the retailer. It is always difficult to forecast sales inbig promotion days, while the existing literature takes those samples asoutliers or focuses on sales forecasting in other industries. Since the O2Ocommerce has its own unique features, such as the combination of both onlineand offline commerce, it is necessary to propose a new method to forecast O2Osales in big promotion days. Therefore, this paper develops a clustering-basedsales forecast method mainly for big promotion days in O2O on-demand retailing.

讲座三：

演讲人↘️：肖沁（凯发平台2017级硕博连读生🧰，技术经济及管理专业）

题目：Item-level Forecasting forTmall Sales with High-dimensional Features via Machine Learning

摘要：With the rapid development ofinformation technology and fast growth of Internet users, e-merchants nowadaysare facing challenges to effectively forecast their demand to allocate theirresources such as inventories and sales efforts. Especially for the item-levelforecasting, the challenges lie on the facts of low amount, high variation, andintermittent demand. Using machine learning techniques such as random forest,this study forecasts daily sales of e-merchants on item-level on Tmall based onhigh-dimensional features, namely, the six feature categories includingcompetition, price, historical sales, page view, online review, and logistics.Our results indicate that the forecast accuracy of random forecast is higherthan the traditional time series method. Among the six feature categories, pageview is the most important in forecasting sales. This study helps e-merchantsforecast their future customer demand using the available online informationand make the corresponding operational decisions.

讲座四🛹：

演讲人🦕：温岩（凯发平台2016级硕博连读生🚡，技术经济及管理专业）

题目：Crowdsourcing, deliveryefficiency, and incentives: An empirical research on the O2O on-demand deliverysystem

摘要🙅：Online-to-offline (O2O)on-demand delivery is characterized by high demand and short time requirements.Traditional employed-driver delivery (i.e., delivery carried out by formallyemployed drivers) cannot fulfill these demands, thereby necessitating thedevelopment of crowd delivery (i.e., delivery carried out by occasionaldrivers) as a new partnership mode. Compared with employed-driver delivery,crowd delivery is more flexible but with lower service quality. It is notobvious which delivery mode outperforms the other in terms of deliveryefficiency. To compare the efficiency of these two delivery modes, this studycollects 131,443 orders from 38 restaurants of an online restaurant chain inChina. Our results show that, compared with employed-driver delivery, crowddelivery increases the order delivery time by an average of two minutes. Thisgap is reduced when the number of orders reaches its peak or delivery pricesare high, but is enlarged during weekends. Our results also suggest that crowddelivery is more suitable for companies that adopt a cost-oriented operationstrategy, while employed-driver delivery is more suitable for companies thatfocus on service quality. Meanwhile, a combination of these two modes fitscompanies that focus on robust delivery. The findings of this research areexpected to further enhance the present understanding of the motivation andefficiency of O2O on-demanddelivery.