時間：2018年4月18日下午15:45

地點：西南交通大學九里校區零號樓0411室

主題一：Joint replenishment decision considering shortages, partial demand substitution, and defective items

主講人：陳彥如，西南交通大學經濟管理學院副教授，博士生導師。加拿大多倫多大學、香港城市大學訪問學者。主要從事運作管理的研究與教學工作。目前主持國家自然科學基金2項、省部級科研項目5項、企業委托項目6項。在“Expert Systems with Applications”、“Transportation Research Part A: Policy and Practice”、“ Computers & Operations Research”、“Computers & Industrial Engineering”等期刊發表論文7篇。合作出版專著教材3部。

摘要：A shortage may occur because of the insufficient production capacity or possible damages of items in transit, and the shortage can be partially fulfilled with substitutable items. In this study, a joint replenishment problem (JRP) with the shortage and partial demand substitution is investigated by developing a mixed integer nonlinear programming model. Several real-world constraints, such as budget, transportation capacity, and shipment requirement constraints are incorporated in the proposed model. Three heuristic algorithms, namely two-dimensional genetic algorithm I, two-dimensional genetic algorithm II, and differential evolution, are proposed, and numerical examples are provided to demonstrate the applicability of the proposed model in a real-world setting. The performance of the heuristic algorithms is investigated with the help of extensive computational experiments. The results show that differential evolution performs best in term of the minimum total cost among three heuristics. Sensitivity analyses are conducted to provide managerial insights. The results indicate that partial demand substitution policy can effectively decrease the total expected cost, but defective items will exponentially increase the total expect cost. The major ordering cost, budget, and truck capacity also affect the system.

主題二：Benchmarking Study for Credit Scoring Model based on Two Datasets

主講人：程賢，西南交通大學經濟管理學院信息系統與運營管理系老師，主要研究方向：網絡相關商務智能，科技金融，大數據金融風險管理。

摘要：Credit scoring, which is concerned with developing empirical models to support decision making in the assessment of credit risk, has attracted significant attention from managers at financial institutions around the world to academic researchers in many related fields, such as personal credit cards, consumer loans, mortgages and P2P lending. Technological advances have increased the febrile state of credit scoring and a large number of technical credit scoring models has emerged.

However, Previous credit scoring modelling literatures has revealed limitations; namely (1), using few or small data sets and (2) using only a small set of conceptually similar performance indicators. Therefore, we perform a systemic benchmarking study for credit scoring models though (1) using two data sets of considerable size and (2) considering several conceptually different performance indicators.

                                     歡迎全院師生參加！