342717022.

xls : Results Page

01/16/2017 13:27:29

Questions Being Addressed by this Spreadsheet

1. What is the expected number of completed cars per day?
2. For a given accumulator size, what is the probability that the downstream segment is starved?
3. For a given accumulator size, what is the probability that the upstream segment is blocked?
4. What are the performance characteristics (MTBF, MTTR, efficiency) of the segments?
Key Assumptions
1. Stations modeled as bernoulli random variables; stations are i.i.d.
2. Segments modeled as binomial random variables.
3. Accumulator is modeled as a M/M/1/c queue where c is the accumulator size+1
4. Shaded cells denote user-specified inputs, unshaded cells denote final or intermediate outputs
Upstream Segment
p(station failure)
E[# cycles until failure]

0.0018
550

# stations/segment
p(segment failure)
p(upstream segment starved)

33
0.0583
0.07

TAKT time (seconds/car)

# hrs/shift
# shifts/day
Maximum # cars/day
MTBF
MTTR
Stand-alone efficiency
Results
Accumulator size
P(downstream seg. starved)
P(upstream seg. blocked)
Expected stock level

100
8
2
576
17.1565
1
0.9449

15
0.08
0.04
16.81

Downstream Segment
p(station failure)
0.0018
E[# cycles until failure]
550

Explanation
Probability that a station experiences a minor failure during a cycle
Expected number of production cycles until a minor failure at station

# stations/segment
p(segment failure)

33
0.0583

The number of stations that comprise a segment

The probability that a segment fails during a cycle
The probability that the upstream segment is starved

103
8
2
559.2233

The cycle time of each station, measured in seconds

The number of hours of production per shift
The number of production shifts per day
The maximum number of cars processed per day

TAKT time (seconds/car)

# hrs/shift
# shifts/day
Maximum # cars/day
MTBF
MTTR
Stand-alone efficiency

17.1565
1
0.9449

Mean number of cycles between segment failures (assumes no idling)

Mean number of cycles required to repair segment
The efficiency of the segment (assumes no idling)

The maximum number of cars that can be held in the accumulator

The probability that the downstream stage is starved
The probability that the upstream stage is blocked
The expected number of cars in the accumulator

Page 1 of 4

342717022.xls : Results Page

during a cycle
failure at station

ssumes no idling)

Page 2 of 4

01/16/2017 13:27:29

342717022.xls : Tradeoff Curves

01/16/2017 13:27:29

Blockage and Starvation Probabilities as a Function of Maximum Accumulator Size

1.0000
0.9000
0.8000
0.7000
Probability

0.6000

P(Downstrea
m starved)
P(Upstream
blocked)

0.5000
0.4000
0.3000
0.2000
0.1000
0.0000
0 1 2 3 4 5 6 7 8 9 1 11 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3
0
2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0
Accumulator Size

3 of 4

342717022.xls : Computations Page

01/16/2017 13:27:30

Key Things To Note:

1. Accumulator is modeled as a M/M/1/c queue where c is the accumulator size+1
2. The arrival rate is computed as Segment 1's arrival rate times (1 - segment 1's failure rate)
3. The departure rate is computed as Segment 2's arrival rate times (1 - segment 2's failure rate)
4. The spreadsheet formulation assumes that the maximum accumulator size is 99

OUTPUTS
Cars in Cars in P(Cars in
Queue System System)
0
0
0.08
0
1
0.49
1
2
0.32
2
3
0.23
3
4
0.18
4
5
0.15
5
6
0.13
6
7
0.11
7
8
0.09
8
9
0.08
9
10
0.07
10
11
0.07
11
12
0.06
12
13
0.05
13
14
0.05
14
15
0.04

INPUTS (from Results Page)

arrival rate
departure rate

0.009
0.009

measured in cars/second
measured in cars/second

rho
Max Accumulator Size
System Size

0.958
15
16

arrival rate / departure rate

user-specified input from Results Page
Has to be accumulator + 1 to include the server

SUMMARY OUTPUTS (to be transferred back to Results Page)

P(Downstream stage starved)

P(Upstream stage blocked)
E[Cars in Queue]
effective arrival rate
E[Cars in System]

0.081
0.041
16.812
0.008
19.338

Page 4 of 4

this is the expected # cars in accumulator

used to determine expected # cars completed
this isn't transferred back; it is artifact of the model