Interactive Tool for an Optimal Equipment Selection in Assembly Lines

1

• Interactive Tool for an Optimal Equipment
  Selection in Assembly Lines
• Collaboration FABRICOM - University of
  Brussels 01/09/95 -
• 31/08/99
• Fabrice Pellichero

2
Contents of the presentation

• Framework of the project
• The equipment pre-selection method
• The automatic resource planning tool
• Conclusions and further works

3
Framework of the project
4
Equipment pre-selection (1)Basic principles

• Based on the use of Functional Groups (FGs)
• Definition Set of equipment able to execute a
  given assembly operation.
• Goal of the pre-selection tool
• propose a set of feasible FGs for each
  operation
• Each FG has an associated
• cost,
• operating time,
• availability.

5
Equipment pre-selection (2) Building of a FG
6
Equipment pre-selection (3) Building of a FG
7
Equipment pre-selection (4) Building of a FG
8
Equipment pre-selection (5) Building of a FG
9
Equipment pre-selection (6) Building of a FG
10
Equipment pre-selection (7)Hidden times
11
Automatic resource planning (1)Input and output
data
Operations FGs (cost, duration, availability)
Number of workstations
Resource - Planning
Operations executed on each workstation
Cycle time
Relative position of each workstation
Precedence constraints
Preference Constraints
FG associated to each operation
SAM
12
Automatic resource planning (2)Genetic
Algorithms (GAs)

• Inspired from evolution of species in Nature
  (objective function acts as environment)
• Maintain a population of solutions
• Work with a representation of the solutions
  (chromosomes)
• New solutions created by combining the best
  members of the population (heredity)
• New solutions replace the worst members of
  the population (survival of the fittest)

13
Automatic resource planning (3)Grouping Genetic
Algorithms (GGAs)

• Special kind of GAs designed to solve grouping
  problems
• work on the groups rather than on the objects
• special encoding scheme
• special operators

14
Automatic resource planning (4) The GGA steps

• 1. Create a population of individuals using the
  Individual Construction Algorithm
• 2. Use the decision-aid method PROMETHEE to
  order individuals in the population (not an
  absolute scalar fitness)
• 3. Recombine (mate) best individuals (parents) to
  produce children (with use of the ICA)
• 4. Mutate children (with use of the ICA)
• 5. Use PROMETHEE to order the new population
• 6. Replace the worst individuals of the
  population by the new children.
• 7. If a satisfactory solution has been found
  stop. Else go to 3

15
Automatic resource planning (5) Individual
Construction Algorithm (ICA)

• 1. Assign tasks (or operations) to the
  workstations (using the operating time
  corresponding to the fastest equipment) according
  to an Equal Piles strategy
• 2. Generate all possible equipment combinations
  for each station thanks to a Branch Cut
  algorithm
• 3. Select the best equipment combination for each
  station using the multicriteria decision-aid
  method PROMETHEE

16
Automatic resource planning (6)Branch and Cut
algorithm

• Enumerates all the possible solutions with the
  help of a search tree
• Cuts the branches of the search tree when it can
  be proven that
• the node does not contain an optimal solution
• the node does not contain a valid solution

17
Automatic resource planning (7)The PROMETHEE II
method

• Multicriteria decision-aid method based on the
  use of
• preference functions associated to each criterion
• weights associated to each criterion
• Computes a net flow ?, which gives a complete
  ranking between the alternatives
• Needs very small computation times while avoiding
  the disadvantage of a complete aggregation

18
Conclusions and further works (1)Advantages of
the resource planning method

• The method allows a real global optimization of
  the line thanks to
• the use of precedence constraints instead of a
  fixed assembly sequence
• the equipment pre-selection tool which proposes
  several FGs by operation
• The method allows a real multicriteria
  optimization thanks to the use of PROMETHEE

19
Conclusions and further works (2)Next
development steps

• Testing on industrial case studies
• Enrichment of the equipment database
• Possibility of automatic link with a simulation
  tool
• Treatment of the multivariant products