Transcript OR/MS in the Networked Economy

OR/MS in the
Networked Economy
Melbourne ASOR AGM, March 2001
Jay Horton
Applied Decision Analysis
PwC
This evening’s presentation
• Some opening remarks
• Background to PricewaterhouseCoopers’ - Applied Decision
Analysis
• How companies are benefiting from OR/MS
• The emergence of Web based Operations Research
PricewaterhouseCoopers
OR/MS in the Networked Economy
Economies of the past:
Network Based Economy:
• “Centralised Allocation” of
goods and services
• “Market Based Allocation” of
goods and services
• Variety of “In-house”
capabilities maintained
• External services utilised as
needed
• Static planning in a stable
environment
• Rapid planning in a dynamic
environment
• Local value chain point of view
• Global value chain prospective
– Sub-optimal relationships
found
– Globally optimal solutions
are feasible
PricewaterhouseCoopers
The need for OR in the Networked
Economy
• OR techniques can manage risk in an objective way
– these days, the present is a poor guide to the future
• Analytical techniques (eg. Data mining) required to extract
useful information from volumes of data
– IT has left us data rich, but information poor
• Modeling business decisions is a fast way to experiment
without risking the performance of a company
– companies need to react quickly and avoid failure
Need for OR Consulting over the Web
PricewaterhouseCoopers
Applied Decision Analysis
• Applied Decision Analysis (ADA) is group within
PricewaterhouseCoopers providing OR/MS skills to help clients
with corporate and financial strategy
• ADA essentially does “Strategic OR”:
– capital and resource allocation
– value chain strategy
– new business initiatives
– market and auction design
PricewaterhouseCoopers
Applied Decision Analysis
• The modelling skills ADA uses:
– Optimisation (LP & MIP), Constraint Programming
– Decision Analysis and Real Option Valuation
– Simulation, Monte-Carlo Analysis
– Game Theory
– Financial Theory
PricewaterhouseCoopers
Modelling Online
• Modelling tools can be
made available to Clients
over the Web when
needed
• Clients access our models
through pwcdecision.com
PricewaterhouseCoopers
How companies are
benefiting from OR/MS
PwC
Tactical Planning for the Coal Mining
Industry
Block 1 Block 2
Block 3
………
Ramp 1
..
Strip &
Mine
Block n
Haul
ROM stockpile
(many more Ramps)
Processor 2
Prod 1
Prod 2
Processor 1
Stockpile
Prod 3
(Many washing
options)
Products
Multi-stage process with many choices at each stage
PricewaterhouseCoopers
MOPS is an optimisation planning tool
for tactical planning
Mine Operations Planning
System (MOPS) is:
• driven by a mixed integer
optimisation model
• a Web based decision
support system
– Allowing multiple users
in various mines
– Administered in Sydney
online
PricewaterhouseCoopers
MOPS provides rapid tactical plans for
operating a mine
MOPS gives:
• Rapid production of plans
• Reduced operating costs
• Product quality ensured
• Proactive risk management
• “What if” analysis capability
PricewaterhouseCoopers
Delivering an integrated manufacturing
solution for the steel industry
• A large producer and distributor of steel, required a strategic
and tactical model to assist in planning production and
distribution activities
• ADA developed a Linear Programming model capable of
optimising its network of plants, processes and its allocation of
products to plants
• Due to the size of the problem, the model needed to be solved
using a high powered computer
Approx. 50
raw materials
12 plant locations
> 20 work centres
Approx. 10,000
products
93 customer
locations
PricewaterhouseCoopers
A major international airline needed a
review of its route allocations
• Matching of routes vs aircraft takes account of many variables
– passenger demands, physical airport limitations,
maintenance requirements
• Too many types of aircraft
– Inefficient maintenance
– Lack of economies of scale
• Financial problems/performance
– Low prices, high costs on parts of the network
PricewaterhouseCoopers
A high-level network optimisation
model was developed
Key Inputs:
Key Outputs:
PAX demand &
fares by class
Optimal allocation
of fleet to routes
Fleet types &
number
Routes by leg
Network
Value
Advisor
Operating costs
Total costs,
revenues and
profits
Shadow prices for
demand and
aircraft hours
Objective: Maximise Profit
Key Constraints:
Must meet demand
Aircraft - route restrictions
Fleet hours available, load factors
PricewaterhouseCoopers
Real Options in 3G Wireless
• A telecommunications firm was considering its future position
with the auction of 3G spectrum
• They had a number of decisions to make:
– how much bandwidth to purchase if any
– how to move forward if bandwidth is not purchased
• They also face a number of uncertainties:
– market demand and price
– future developments in technology
PricewaterhouseCoopers
Real Options in 3G Wireless
• ADA used Real Options Valuation to evaluate the various future
paths and to guide decision making
Spectrum
Cost
Success in
New Market
Below estimate
At estimate
Sales Potential
in New Market
Minimal
Baseline
Revenue
Growth
Early
Product
Expansion
Unsuccessful
Low
Moderate
Nominal
High
Yes
No
Yes
No
Yes
No
Maintain original plan
Highly Successful
Above estimate
Moderate
Unsuccessful
Low
High
Moderate
Nominal
High
Highly Successful
Yes
No
Yes
No
Yes
No
Expand early when
initial indications are
that sales potential is
high
PricewaterhouseCoopers
Bidder support in spectrum auctions
• There is inherent uncertainty in
the auction
• Risk of serious exposure
problems
– Insufficient spectrum
bandwidth
– Insufficient geographic
scope
• Rules are complex
PricewaterhouseCoopers
The emergence of Web
based Decision Support
PwC
Why Web based Decision Support?
A Web based approach to decision support:
• Facilitates a Global focus and rapid planning
• Creates economies of scale in:
– expertise
– computing power
• Allows decentralised use, with less hassle
• Allows better support, on-line
PricewaterhouseCoopers
The Web facilitates global planning
Client
Model
library
Industry
data
base
Consultant
The Web allows global access to the decision support:
– On-line models
– Consulting advice
– Links to industry data
this means planning can be conducted very rapidly
PricewaterhouseCoopers
Creating economies of scale in decision
support
• Powerful hardware and software is usually required for
Decision Support Systems
• Putting a system online allows the centralisation of computing
power
• With many users, scale is gained. This:
– lowers the cost to users
– increases the model solving capability
PricewaterhouseCoopers
‘Over time, operations research, enterprise systems,
advanced planning and scheduling, customer
relationship management and the internet will
gradually become more and more blended. This
phenomenon is now changing the meaning of the
word, operations research, itself.’
PricewaterhouseCoopers