]ISysE / GSDS 세미나가 다음과 같이 진행될 예정입니다.

# 날짜/시간: 2024년 6월 3일 월요일 9:00~10:00

# 장소 : E-2동 2122호

# 연사 : Professor Suresh Sethi / Operations Management / University of Texas at Dallas

# 제목 : Optimality of Base Stock Policy  under Unknown General Demand  Distributions:  New Methods and New Results 

# 초록 :  This paper advances the literature on the  optimality of the base stock policy for a general  demand distribution, and a general prior belief,  which we update as we observe realized  demands, assumed to be continuous, i.i.d.,  random variables. The value function depends  on the belief, so the functional Bellman  equation is infinite-dimensional. Significantly, in  contrast with the traditional approach, we  derive a functional equation for the derivative  of the value function with respect to the  inventory level, which provides a direct  approach to obtaining the optimal base stock.  In two well-known cases, we characterize how  the base stock level depends on the belief, and  we implement the approach to obtain the  optimal base stock. In the case of conjugate  probabilities, the infinite-dimensional state  reduces to a finite-dimensional sufficient  statistic. That allows us to solve a numerical  example of Weibull demand. The second case  considers the demand to come from one of two  possible distributions, but we do not know  which. Here, we derive a functional equation in  one hyperparameter expressing the ratio of the  weights assigned to the two distributions. We  then develop an approximation scheme to  solve it, show that it converges, and implement  it numerically to obtain the optimal base stock.

 

# 본 세미나는 영어로 진행됩니다.

 

 

---------------------------------------------------------------------------------------------------------------------------------

 

ISysE/GSDS dept. office invites you to the following seminar.

 

# Time/Date : June 3, 2024 (Mon.) 9:00~10:00

# Location : Room 2122 in E-2 building 

# Presenter : Professor Suresh Sethi / Operations Management / University of Texas at Dallas

# Title : Optimality of Base Stock Policy  under Unknown General Demand  Distributions:  New Methods and New Results 

# Abstract :  This paper advances the literature on the  optimality of the base stock policy for a general  demand distribution, and a general prior belief,  which we update as we observe realized  demands, assumed to be continuous, i.i.d.,  rando variables. The value function depends  on the belief, so the functional Bellman  equation is infinite-dimensional. Significantly, in  contrast with the traditional approach, we  derive a functional equation for the derivative  of the value function with respect to the  inventory level, which provides a direct  approach to obtaining the optimal base stock.  In two well-known cases, we characterize how  the base stock level depends on the belief, and  we implement the approach to obtain the  optimal base stock. In the case of conjugate  probabilities, the infinite-dimensional state  reduces to a finite-dimensional sufficient  statistic. That allows us to solve a numerical  example of Weibull demand. The second case  considers the demand to come from one of two  possible distributions, but we do not know  which. Here, we derive a functional equation in  one hyperparameter expressing the ratio of the  weights assigned to the two distributions. We  then develop an approximation scheme to  solve it, show that it converges, and implement  it numerically to obtain the optimal base stock.

 

# The seminar will be in English