"prescriptive analytics for logistic management โดย...
TRANSCRIPT
Prescriptive Analytics for Logistics Management
The First NIDA Business Analytics and Data Sciences Contest/Conferenceวันที่ 1-2 กันยายน 2559 ณ อาคารนวมินทราธิราช สถาบันบัณฑิตพัฒนบริหารศาสตร์
•What is Prescriptive Analytics? •Prescriptive Analytics & Competitive advantage •Case study: Making “Better” Decisions in Logistics Management using Prescriptive Analytics
https://businessanalyticsnida.wordpress.comhttps://www.facebook.com/BusinessAnalyticsNIDA/
รศ.ดร.กาญจ์นภา อมรัชกุลสาขาวิชา Logistic Managementคณะสถิติประยุกต์ NIDA
นวมินทราธิราช 3003วันที่ 2 กันยายน 2559 10.45-11.15 น.
PRESCRIPTIVE ANALYTICS FOR LOGISTICS MANAGEMENT
Kannapha Amaruchkul
Logistics Management Program,
School of Applied Statistics, NIDA
2016-09-02 10:45-11:15 Rm3003
ANALYTIC VALUE ESCALATOR
3
PREDICTIVE VS PRESCRIPTIVE ANALYTICS
Predictive Analytics
o Show trends for the KPIs
o Create spot opportunities
1. Revise Estimated Time of Arrival (ETA) based on specific driver history and real-time traffic
2. What will happen to revenue if 30% discount is given to customer in segment A123?
3. What will happen if we keep 2-week of inventory?
Prescriptive Analytics
o Determine where your KPIs should be
o “Prescribe” how to get there
1. Optimize route to minimize average travel time
2. What is the best discount plan that would maximize the total revenue?
3. What is the optimal inventory policy that minimizes the expected total cost subject to pre-specified customer service level?
4
http://blogs-images.forbes.com/louiscolumbus/files/2015/07/Figure-1-SCM-Data-Volume-Velocity-Variety.jpg5
LOGISTICS NETWORK (SUPPLY CHAIN)
6Source: Simchi-Levi, D., & Kaminsky, P., & Simchi-Levi, E. (2007). Designing and Managing the Supply Chain: Concepts, Strategies and Case Studies. Boston: McGraw-Hill.
WHAT MAKES SUPPLY CHAIN MANAGEMENT DIFFICULT?
7
Global optimization:
It is challenging to operate supply chain so that total systemwide costs are minimized, while systemwide service levels are maintained.
Conflicting objective among players
The process of finding the best systemwide strategy is referred to as global optimization.
Managing uncertainties:
In every supply chain, “the only thing certain is uncertainty.”
Customer demand will not be known in advanced, and machines will break down at some point in time.
Supply chain management needs to deal effectively with uncertainties.
KEY ISSUES IN SUPPLY CHAIN MANAGEMENT
Global optimization Managing risk and
uncertainty
Distribution network configuration
Inventory control
Production sourcing
Distribution strategies
Supply contracts
Strategic partnering
Outsourcing and offshoring
Product design
Information technology
Smart pricing8
COMPETING ON ANALYTICS
9
Com
petitive A
dvanta
ge Stochastic Optimization
How can we achieve the best outcome taking account of
uncertaintyPrescriptive
Optimization How can we achieve the best outcome?
Predictive modeling What will happen next if…?
PredictiveForecasting What if these trends continue?
Simulation What could happen…?
Alerts What actions are needed?
Query What exactly is the problem?
DescriptiveAd hoc reporting How many, how often, where?
Standard reporting What happened?
Degree of complexity Davenport & Harris (2007)
EXAMPLES OF OPTIMIZATION MODELS
Model context Decision variables Objective function Typical constraints
Product mix Quantities of product to produce Maximize profit Resource limitations (e.g.,
production time, labor,
material); maximum sales
potential
Blending Quantity of materials to mix to
produce one unit of output
Minimize cost Demand requirements; resource
limitation
Transportation Amount to ship between sources of
supply and destination
Minimize total transportation
cost
Supply constraints; require
demands met at destinations
Multi-period production
planning
Quantities of product to produce in
each several tie periods; amount of
inventory to hold between periods
Minimize total production and
inventory costs
Limited production rates;
material balance equations
Marketing Allocation of advertising
expenditures; production quantities
Maximize profit Budget limitation; production
limitations; demand
requirements
10
STOCHASTIC MODELS
Markov chain
Customer relationship management
Inventory control
Accounts receivable
Warranty policy
Queueing models
Transportation management
Warehouse management
Markov decision process (MDP)
Equipment replacement model
Lot sizing model (Wagner-Whitin algorithm)
11