Operation management

Disponível somente no TrabalhosFeitos
  • Páginas : 8 (1792 palavras )
  • Download(s) : 0
  • Publicado : 22 de janeiro de 2015
Ler documento completo
Amostra do texto
“VNS for glass container industry
Capacitated Lotsizing problem”

B Almada-Lobo1; MA Caravilla1,2; JF Oliveira1,2
1
2

FEUP – Faculdade de Engenharia da Universidade do Porto
INESC Porto – Instituto de Engenharia e Sistemas de Computadores

almada.lobo@fe.up.pt

MIC 2007, 25th June

MIC 2007

Agenda
1. Glass Container Manufacturing Process
2. Glass Container Industry
3.Long-Term Production Planning
4. A Solution Approach: VNS
5. Conclusions

Bernardo Almada-Lobo

2

MIC 2007

1 Glass Container Manufacturing Process
Three main sub-processes
-

Glass Production

-

Containers Manufacturing

-

Palletising

Raw materials
silos

Bernardo Almada-Lobo

One supporting sub-process
-

Hot-end

Decoration

Cold-end
Quality Control

3Final Product

MIC 2007

1 Glass Container Manufacturing Process

Reception and Control of Raw Materials
Raw Materials Storage
Filling, Weighting and Mixing

Bernardo Almada-Lobo

4

MIC 2007

1 Glass Container Manufacturing Process
Furnace Melting Capacity (tonnes/day) → f (Colour)

Parallel Moulding Machines

Furnaces operate continuously

Moulding machine has four maincharacteristics:

Furnace output can vary to a limited extent
high sequence dependent setup times involved in a
colour change → furnace colour specialization
Glass paste colour (amber, flint, green)

-

Number of individual sections

-

Number of mould cavities per section (single,
double or triple-gob machine)

-

Distance between the moulds

-

Manufacturing processBlow-blow
Press-and-blow

Not negligible sequence dependent setup times
in product changeovers
‘Machine balancing’
Job-splitting not allowed

Bernardo Almada-Lobo

5

MIC 2007

1 Glass Container Manufacturing Process

Inspection
Shipping

Reheating Kiln

Palletising

Surface treatment

Bernardo Almada-Lobo

6

MIC 2007

1 Glass Container Manufacturing ProcessRaw-materials

Melting rate Furnace
(tonnes/day)
Hot zone
Feeders

M1

M2

(⋅ ⋅ ⋅)

Mn

Molding Machines
(with different settings)

Lehrs

Cold zone
Inspection

Packing

Bernardo Almada-Lobo

7

Semi-Continuous process
Single-Operation type

MIC 2007

2 Glass Container Industry
Sales of glass containers:
-

High seasonality (beer in Summer…)
High variability(intermediary product)
Make-to-Stock Industry

Production capacity almost constant
Changeover costs
Container glass is typically manufactured closer to the end user
Strong focus on improving efficiencies and reducing costs to
remain competitive
Capital intensive industry

Bernardo Almada-Lobo

8

MIC 2007

3 Long Term Production Planning
3.1

Objectives

Maximize met demand

Maximizefacilities’ throughput

Minimize inventory levels

Bernardo Almada-Lobo

9

MIC 2007

3 Long Term Production Planning
3.2

Output

Output: 12 month rolling horizon plan

Bernardo Almada-Lobo

-

Colour campaign scheduling

-

Assignment of products to machines monthly

10

MIC 2007

3 Long Term Production Planning
3.3

Mathematical Model

MILP Formulation
-Extension of Single-machine CLSP with
sequence dependent setups [AlmadaLobo, 2007]

-

Additional constraints

-

Parallel machine environment + sequence
dependent setup times

CLSP is NP-hard [Bitran and Yanasse,
1982]
-

Finding a feasible solution is NP-complete
[Maes et al, 1991]
Size of our
instances

Heuristic Procedures to solve this
problem
Bernardo Almada-Lobo11

MIC 2007

4 A Solution approach
4.1

VNS

Neighbourhood changes to avoid a local optimum entrapment.

Basic VNS: systematic search through increasingly distant
neighbourhoods
-

Which set of NS must be used in the shaking phase?

-

Which set in the local search phase?

-

How to order those neighbourhoods?

Bernardo Almada-Lobo

12

MIC 2007

4 A Solution...
tracking img