Operational

1
Operational Research (O.R.) Techniques
Simulation
2
Supermarket checkouts

• Imagine that you need to decide on the checkout
  arrangements at a supermarket
• How many checkouts?
• How many basket only less than 5 items, less
  than 10?
• How many self service?
• What would you do?

3
You could approach this problem in two ways

• Experiment directly
• This is expensive and time consuming and may not
  give accurate results as the different layouts
  will be trialled at different times and not under
  the same circumstances.
• Build a model
• A computer simulation model allows different
  layouts to be run while all other elements remain
  the same. The differences between the results
  will then be due to the different arrangements of
  the checkouts.

4
Simulation

• Simulation is a technique used extensively in
  business to aid the planning of new operations
  and investigate better ways of working current
  systems.
• Different types of simulation
• Discrete
• The variables of interest change only at
  particular instants of time, e.g. train arrives,
  the doors open, the train departs.
• This type of simulation is known as Discrete
  Event Simulation.
• Continuous
• The variables of interest are changing
  continuously with time, e.g. the speed of a
  train.
• This type of simulation is tackled using a
  methodology called System Dynamics.
•  

5
Why simulate?

• Cost simulation may be time consuming and
  expensive in terms of skilled man power but if a
  real experiment goes wrong it could be very
  expensive.
• Time it may take time to produce programmes for
  simulation but then it is possible to simulate
  weeks, months even years in seconds of computer
  time.
• Replication simulation allows repetition which
  means a whole range of policies can be
  implemented without affecting the real world.
• Safety the effect of extreme conditions may be
  one objective of simulation, to do this in real
  life may be dangerous or illegal.

6
SIMUL8

• SIMUL8 simulation software is used to model
  discrete event simulations.
• Petrol station model

7
Petrol station model

• Model structure
• Cars arrive
• Wait for petrol pump to become free
• Use the pump
• Queue to pay
• Leave

8
Petrol station model

• Run base case scenario.
• What was the average queue size that was waiting
  for fuel?
• What was the maximum time spent queuing waiting
  for fuel?
• What was the maximum queue size waiting to pay?

9
Petrol station model

• Parameters that can be changed in the model
• Number of cars per hour
• Number of petrol pumps
• Number of servers at pay

10
Petrol station model
11
Petrol station model

• What is the effect of changing the number of cars
  per hour?
• What effect does changing the other parameters
  have?
• What is the most effective combination for
  minimising the queues and maximising the number
  of customers?

12
STELLA

• STELLA simulation software is used to model
  system dynamics or continuous simulations.
• H1N1 Flu Outbreak model

13
Flu outbreak model

• The core model structure of this system is a main
  chain of the student population which flows
  through the stocks of Susceptible, Exposed,
  Infected and Recovered.
• Students are susceptible to the virus.
• They then become exposed when come into contact
  with an ill student.
• A few days after exposure the symptoms are
  visible and the student is infected.
• After a number of days the infected student gets
  better and is recovered.

14
Flu outbreak model

• Run base case scenario
• How many students become sick?
• Variables that can be changed in the model
• vaccinated
• effectiveness of vaccine
• Average number of days infected students stay at
  home

15
Flu outbreak model
16
Flu outbreak model

• What is the effect of changing the percentage of
  students vaccinated?
• What effect does changing the other parameters
  have?
• What is the most effective combination to
  minimise the outbreak of the flu in this school?