A Multidisciplinary Approach

1
A Multidisciplinary Approach to Enterprise-wide
Optimization Ignacio E. Grossmann Carnegie
Mellon University Pittsburgh, PA 15213 Jeffrey
Linderoth Lehigh University Bethlehem, PA 18015
2
Enterprise-wide Optimization

• The supply chain is large, complex, and highly
  dynamic
• Optimization can have very large financial payout

3
Pharmaceutical supply chain
(Gardner et al , 2003)

• Pharmaceutical process (Shah, 2003)
• Primary production has five synthesis stages
• Two secondary manufacturing sites
• Global market

4
Enterprise-wide Optimization (EWO)
EWO involves optimizing the operations of RD,
material supply, manufacturing, distribution and
financial activities of a company to reduce costs
and inventories, and to maximize profits, asset
utilization, responsiveness and customer
satisfaction.

• Key features
• Integration of the information, modeling and
  solution methods

5
Integration IT/Methods for Enterprise-wide
Optimization
6
Research Challenges

• The modeling challenge
• Planning, scheduling, control models for the
  various components of the supply chain, including
  nonlinear process models?

Novel math. programming, logic-based, dynamic
models
- The multi-scale optimization challenge
Coordinated optimization of models over
geographically distributed sites, and over the
long-term (years), medium-term (months) and
short-term (days, min) decisions?
Novel decomposition procedures across large
spatial and temporal scales.
7
Research Challenges
- The uncertainty challenge How to effectively
anticipate effect of uncertainties ?
Novel and meaningful stochastic programming tools.
- Algorithmic and computational challenges How
to effectively solve the very-large scale models
in terms of efficient algorithms, and in terms of
modern computer architectures?
Novel algorithms, implemented in distributed or
grid computing.
8
Multidisciplinary team Enterprise-wide
Optimization Chemical engineers, Operations
Research, Industrial Engineering
Researchers Carnegie Mellon Ignacio
Grossmann (ChE) Larry Biegler (ChE)
John Hooker (OR) Lehigh University
Jeffrey Linderoth (Ind. Eng) Univ. Pittsburgh
Andrew Schaeffer (Ind. Eng.)
9
Overall Goals
1. Develop comprehensive set of computational
capabilities for Enterprise-wide
Optimization

• Novel planning and scheduling models, including
  consideration of uncertainty
• Effective integration of Production Planning,
  Scheduling and Realtime Optimization
• Optimization of Entire Supply Chains

10
Overall Goals (cont.)
2. Test methods with case studies from industry
Group Enterprise-wide Optimization for
Process Industries
Case studies with partner companies ABB Crane
scheduling steel manufacturing Contact Iiro
Harjunkoski John Hooker Air Products Liquid
Bulk-Gas Production Distribution Contact Ken
Anselmo Jeff Linderoth BP Optimization of
multiperiod MILP production model PX and
PTA Contact Norm Jerome Andrew Schaefer Dow
Simultaneous Tactical Planning and Production
Scheduling of Batch Operated Reactors
Contact John Wassick Ignacio Grossmann ExxonMobil
Global optimization of MINLP models for
refinery/petroleum applications Contact Kevin
Furman Ignacio Grossmann NOVA Chemicals
Polymerization optimization Contact Paul
Arch Larry Biegler
11
Approach Dynamic programming models combined
with local search methods
12
(No Transcript)
13
Approach Extension of deterministic planning
model to various form of Stochastic programming
models to handle demands in uncertainties
14
Production Planning and Scheduling Parallel Batch
Reactors Sequence dependent changeovers
DETERMINE PRODUCTION PLAN and SCHEDULE

• Production quantities
• Inventory levels
• Number of batches of each product
• Assignments of products to available processing
  equipment
• Detailed timing of operations
• Sequence of production in each processing
  equipment

OBJECTIVE
To Maximize Profit Sales Costs CostsOperating
Costs Inventory Costs Transition Costs

Approach Novel MILP Planning Model that
anticipates effects of changeover times
Accurate Planning Model integrated with
Scheduling Model
15
- Scheduling and Planning of flow of crude oil is
key problem in petrochemical refineries - Large
cost savings can be realized with an optimum
schedule for the movement of crude oil
Storage
Charging
Vessels
Crude-Distillation Unit
How to coordinate discharge of vessels with
loading to storage? How to synchronize charging
tanks with crude-oil distillation?
Approach Global MINLP Scheduling Optimization
Model in Continuous Time
16

• Packaging design optimization
• Given
• Size and weight of article
• Value of article
• Fragility of article
• Shipping mode controlled vs. uncontrolled
• Distance
• Goal Minimize Cost
• Transportation
• Packaging
• Damaged goods returned
• w.r.t.
• Cushioning material selection
• Foam cushion design
• Foam density (expansion factor)
• Foam thickness
• Details of foam package design ribs, use FEA,
  etc.

17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
http//egon.cheme.cmu.edu/ewocp/
23
EWO Seminars Stochastic Programming Andrew
Schaefer Batch Scheduling Models Ignacio
Grossmann Real-time Optimization Larry
Biegler Constraint Programming John
Hooker Grid Computing Jeff Linderoth Hybrid
Methods for Planning and Scheduling John Hooker
Markov Decision Processes Andrew
Schaefer Overview of MILP Jeff Linderoth
Overview of MINLP Ignacio Grossmann
Dynamic Real-Time Optimization Larry
Biegler
24

• Concluding remarks
• Enterprise-wide Optimization great practical
  relevance, but
• currently beyond the reach of current
  computational capabilities
• Area requires multidisciplinary approach
• Chemical engineers
• Computer scientists
• Industrial engineers
• Operations researchers
• 3. Fundamental challenges include modeling,
  multiscale, uncertainty,
• algorithmic issues
• 4. Grid computing has the potential of supporting
  computational
• challenges that arise in this area
• a key component of cyberinfrastructure