project crashing presented by: meysam rahimi. problem statement: a town council wishes to construct...

29
Project crashing Presented by: Meysam rahimi

Upload: sheila-simon

Post on 03-Jan-2016

226 views

Category:

Documents


5 download

TRANSCRIPT

Page 1: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Project crashing

Presented by:Meysam rahimi

Page 2: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Problem statement:

• A town council wishes to construct a small stadium in order to improve the services provided to the people living in the district. After the invitation to tender, a local construction company is awarded the contract and wishes to complete the task within the shortest possible time. All the major tasks are listed in the following table. The durations are expressed in weeks. Some tasks can only start after the completion of certain other tasks.

Page 3: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Problem statement:The town council would like the project to terminate earlier than the time announced by the builder To obtain this, the council is prepared to pay a bonus of BC 30K for every week the work finishes early. The builder needs to employ additional workers and rent more equipment to cut down on the total time. In the preceding table he has summarized the maximum number of weeks he can save per task (column "Max. reduct.") and the associated additional cost per week.

Page 4: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Problem Data:

Page 5: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Precedence graph of construction tasks:

Page 6: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Two important questions:

1. Which is the earliest possible date of completing the construction?

2. When will the project be completed if the builder wishes to maximize his profit?

Page 7: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Model formulation for question 1:

Page 8: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Critical path method:

Page 9: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Model formulation for question 2:

Page 10: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Crashing algorithm:

• To shorten a project, crash only activities that are critical.

• Crash from least expensive to most expensive.

• Each activity can be crashed until-it reaches it’s maximum time reduction-it causes another path to also become critical-it is more expensive to crash than not to crash

• Continue until no more activities should be crashed.

Page 11: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Example:

• Bonus=$1400

Page 12: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Example cont.

• ABD 18• ACD 19• ACE 20

Page 13: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Example cont.

• E is least expensive to crash.• we save $1400 per day the project is

shortened and would spend $700 per day to crash E, so it is profitable to crash E.

• E has maximum time reduction of 3, but if it is crashed by 1, then ACD also becomes a critical path.

• So we should crash E by 1 period.

Page 14: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Example cont.

• ABD 18 18• ACD 19 19 *• ACE 20 * 19 *

Page 15: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Example cont.

• Now we have 3 choices to crash:– Crash A with the cost of $1000– Crash C with the cost of $2500– Crash both D&E with the cost of $3700

• Crashing A is less expensive and laso is profitable because of being less than $1400.

Page 16: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Example cont.

• We can reduce A to its minimum duration, because no new task will become critical by the crashing of A.

Page 17: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Example cont.

• To continue, we could crash C or both D and E. But in each case, the cost would be greater than the $1400 savings per day. So, we stop at this point.

Page 18: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Uncertainty in activity durations:

• In the real word, we may encounter with uncertainty in activity durations.

• There are three main approaches to deal with uncertainty in activity duration:– Fuzzy logic– Robust optimization– Stochastic optimization

Page 19: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Project scheduling,stochastic approach:

Page 20: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Distribution of completion time:

Page 21: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Expected completion time approximation method:

• Traditional method:– PERT(Program evaluation and review

technique)• Converts the stochastic model to a

deterministic model• Then uses CPM to solve it.• Is a very optimistic method

– Upper and lower bounds

• Monte Carlo simulation

Page 22: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Monte Carlo simulation:

• Generate random activity durations• Find the longest path by CPM• the average of completion time in

samples is an unbiased estimator for the expected value of completion time.

• The standard deviation of sample is also an unbiased estimator for

Page 23: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Distribution of completion timeexample:

U(5,7)

U(4,8)

U(10,16)U(4,6)

U(7 ,9)U(5,9)

U(12,16)

Page 24: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Simulation results:

PERT estimation:21Simulation estimation: 21.68

Page 25: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Activity crashing in stochastic networks:

Page 26: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Activity crashing in stochastic networks:• Because we don’t have the distribution of

completion time, we may get into trouble with the problem.

• All the methods which has been presented to solve this problem are heuristic methods– most critical first– Least expensive first– Combination of two above methods– BBI

Page 27: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Optimization with embedded simulation:• Can be used with improvement

algorithms– Steepest descent – Newton method– SQP

• Uses simulation for evaluation of objective function and finding the search direction for the next iteration.

Page 28: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Optimization with embedded simulation:

Start with a feasible point

SimulationEvaluation of current

solution

OptimizationFinding next solution

Page 29: Project crashing Presented by: Meysam rahimi. Problem statement: A town council wishes to construct a small stadium in order to improve the services provided

Any questions?