COMMERCE 2OC3 Study Guide - Final Guide: Probability Density Function, Random Variable, Probability Distribution
5 Jun 2018
School
Department
Course
Professor
Scientific advances (economics, statistics, CS, OM) allow for greater accuracy in models
•
Analytics and big data allow for greater storage and analysis
•
Yield management is of interest when the service or product can be sold in advance of consumption, demand
fluctuates, resource capacity is relatively fixed, demand can be segmented, low variable costs, high fixed costs
•
Revenue management is the aggregate planning process of allocating scarce resources to customers to maximize
revenue, consisting of price and quantity decisions.
Multiple pricing structures must be feasible and justifiable to a customer to create customer segments with
price levels
•
Forecasts on use and duration as well as changes in demand must be known to prevent overbooking
•
Price is fixed
Price is variable
Use is predictable
Movies, stadiums, convention centers, hotel meeting
spaces
Hotels, airlines, rental cars, cruise
lines
Use is uncertain
Restaurants, golf courses, internet service providers
Hospitals, continuing car
Single-Item Revenue Management
This focuses perishable products with uncertain demand. Approach using the newsvendor model.
If demand is deterministic, use linear or integer programming
○
If demand is uncertain, use stochastic dynamic programming
○
When revenue is optimized, either one or several prices can be defined
•
Critical fractile and optimal quantity
Attributes of the Queueing System
Queueing theory is the quantitative study of queueing systems focusing on uncertainty in the arrival and service
processes in a long-run analysis.
Arrival
characteristic
Description
Examples
Size of arrival
population
Finite population in a queue with
limited number of potential users.
Infinite population in a queue with
unlimited number of potential users.
Car wash, supermarket, student registration
Copying shop with limited copiers
Pattern of
arrivals
Determined using Poisson distribution,
a discrete probability distribution that
often describes the arrival rate in
queueing theory.
= probability of x arrivals
= number of arrivals per unit of time
= average arrival rate
= Euler's constant
Behaviour of
arrivals
Balking customers refuse to enter queue, reneging customers leave when they become impatient.
Waiting line characteristic
Description
Examples
Queue discipline
FIFO describes "first in, first out."
Priority treatment pre-empts FIFO.
Checkout line
Emergency room, express checkout lines
Service
characteristics
Description
Examples
Design of
service system
Single-channel has one line and one server.
Multi-channel has one line and several servers.
Single-phase has one station while multi-phase has multiple.
Dentist office
McDonald's drive through
Post office | College registration
Distribution of service times
Described by negative exponential probability distribution, a continuous distribution.
Performance Characteristics of a Queuing System
Revenue Management & Queueing
March 23, 2018
2:37 PM
Operations Management Page 1
Document Summary
Revenue management is the aggregate planning process of allocating scarce resources to customers to maximize revenue, consisting of price and quantity decisions. Scientific advances (economics, statistics, cs, om) allow for greater accuracy in models. Analytics and big data allow for greater storage and analysis. Yield management is of interest when the service or product can be sold in advance of consumption, demand fluctuates, resource capacity is relatively fixed, demand can be segmented, low variable costs, high fixed costs. Use is predictable movies, stadiums, convention centers, hotel meeting spaces. Multiple pricing structures must be feasible and justifiable to a customer to create customer segments with price levels. Forecasts on use and duration as well as changes in demand must be known to prevent overbooking. When revenue is optimized, either one or several prices can be defined. If demand is deterministic, use linear or integer programming. If demand is uncertain, use stochastic dynamic programming.
