PRODUCTION PLANNING AND

CONTROL
ASSOC. PROF. PERAL TOKTAŞ-PALUT

Lecture 11
Agenda

 Material requirements planning (cont.’d)

 Alternative lot sizing schemes (cont.’d)
◼ The Silver-Meal heuristic
◼ Least unit cost heuristic
◼ Part-period balancing heuristic
 Example: Heuristic methods

 Complete the MRP record for item A.

Lot Size: Best heuristic method Lead Time: 2 weeks

Item: A
Weeks
1 2 3 4 5 6 7 8 9 10 11 12
Gross requirements 130 70 90 70 30 100 60 40 80 20
Net requirements 50 60 90 70 30 100 60 40 80 20
Scheduled receipts 10
Projected on hand
80 80 80
inventory
Planned receipts
Planned order
releases
The Silver-Meal heuristic
 Silver and Meal (1973)
 A heuristic that gives an approximate solution
 A sequential method
 When determining the delivery in period 1, we consider
successively the demands in periods 2, 3, and so on.
 When considering period 2, we use a simple test to decide
whether this period demand should be added to the delivery
batch in the first period. If this is the case we continue with
period 3, etc.
 Assume that this procedure shows that the next delivery will
come in period k. The procedure is then repeated with period k
as the “first period”.
The Silver-Meal heuristic

 To choose to have a new delivery when the average per

period costs increase for the first time
 The first delivery shall cover n periods and the next
delivery take place in period n+1 if
k k −1
S + h ( j − 1 ) d j S + h ( j − 1 ) d j
j =2 j =2
 , 2k n
k k −1
and
n+1 n
S + h ( j − 1 ) d j S + h ( j − 1 ) d j
j =2 j =2

n +1 n
Example 1: The Silver-Meal heuristic

 Starting from week 3, the demands (net requirements) in

T = 10 weeks

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20

 Ordering cost S = $300

 Holding cost h = $1 per unit and week
Example 1: The Silver-Meal heuristic

 If the delivery in period 1 covers only the demand in

period 1, the cost for this period
 S = 300
Example 1: The Silver-Meal heuristic

 We now successively consider the per period costs for

deliveries that cover 2, 3,. . . periods
 2 periods ( 300 + 60 ) 2 = 180  300

 3 periods ( 300 + 60 + 2  90 ) 3 = 180  180

 4 periods ( 300 + 60 + 2  90 + 3  70 ) 4 = 187.5  180

 A new delivery in period 4

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 1: The Silver-Meal heuristic

 The same procedure is applied with period 4 as the first

period
 2 periods ( 300 + 30 ) 2 = 165  300

 3 periods ( 300 + 30 + 2  100 ) 3 = 176.67  165

 A new delivery in period 6

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 1: The Silver-Meal heuristic

 The same procedure is applied with period 6 as the first

period
 2 periods ( 300 + 60 ) 2 = 180  300

 3 periods ( 300 + 60 + 2  40 ) 3 = 146.67  180

 4 periods ( 300 + 60 + 2  40 + 3  80 ) 4 = 170  146.67

 A new delivery in period 9

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 1: The Silver-Meal heuristic

 The same procedure is applied with period 9 as the first

period
 2 periods ( 300 + 20 ) 2 = 160  300

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 1: The Silver-Meal heuristic

 The solution
 A delivery in period 1
 A new delivery in period 4

 A new delivery in period 6

 A new delivery in period 9

Period t 1 2 3 4 5 6 7 8 9 10
dt 50 60 90 70 30 100 60 40 80 20
Solution 200 100 200 100
Example 1: The Silver-Meal heuristic

 Total cost of the solution

 Ordering cost = 4  300 = 1200
 Holding cost = 60 + (2  90) + 30 + 60 + (2  40) + 20 = 430

 Total cost = 1630

 5% higher than the optimal costs obtained by the Wagner-

Whitin algorithm

Period t 1 2 3 4 5 6 7 8 9 10
dt 50 60 90 70 30 100 60 40 80 20
Solution 200 100 200 100
Least unit cost heuristic

 Identical to the Silver-Meal heuristic except that it

accumulates requirements until the cost per unit
increases
Example 2: Least unit cost heuristic

 Starting from week 3, the demands (net requirements) in

T = 10 weeks

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20

 Ordering cost S = $300

 Holding cost h = $1 per unit and week
Example 2: Least unit cost heuristic

 If the delivery in period 1 covers only the demand in

period 1, the total cost per unit
 300 / 50 = 6.00
Example 2: Least unit cost heuristic

 We now successively consider the total costs per unit for

deliveries that cover 2, 3,. . . periods
 2 periods ( 300 + 60 ) 110 = 3.27  6.00

 3 periods ( 300 + 60 + 2  90 ) 200 = 2.70  3.27

 4 periods ( 300 + 60 + 2  90 + 3  70 ) 270 = 2.78  2.70

 A new delivery in period 4

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 2: Least unit cost heuristic

 The same procedure is applied with period 4 as the first

period
 1 period 300 70 = 4.29

 2 periods ( 300 + 30 ) 100 = 3.30  4.29

 3 periods ( 300 + 30 + 2  100 ) 200 = 2.65  3.30

 4 periods ( 300 + 30 + 2  100 + 3  60 ) 260 = 2.73  2.65

 A new delivery in period 7

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 2: Least unit cost heuristic

 The same procedure is applied with period 7 as the first

period
 1 period 300 60 = 5.00

 2 periods ( 300 + 40 ) 100 = 3.40  5.00

 3 periods ( 300 + 40 + 2  80 ) 180 = 2.78  3.40

 4 periods ( 300 + 40 + 2  80 + 3  20 ) 200 = 2.80  2.78

 A new delivery in period 10

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 2: Least unit cost heuristic

 The solution
 A delivery in period 1
 A new delivery in period 4

 A new delivery in period 7

 A new delivery in period 10

Period t 1 2 3 4 5 6 7 8 9 10
dt 50 60 90 70 30 100 60 40 80 20
Solution 200 200 180 20
Example 2: Least unit cost heuristic

 Total cost of the solution

 Ordering cost = 4  300 = 1200
 Holding cost = 60 + (2  90) + 30 + (2  100) + 40
+ (2  80) = 670
 Total cost = 1870

 20% higher than the optimal costs obtained by the Wagner-

Whitin algorithm

Period t 1 2 3 4 5 6 7 8 9 10
dt 50 60 90 70 30 100 60 40 80 20
Solution 200 200 180 20
Part-period balancing heuristic

 To select the number of periods covered by the

replenishment such that the total holding costs are
made as close as possible to the ordering cost
Example 3: Part-period balancing heuristic

 Starting from week 3, the demands (net requirements) in

T = 10 weeks

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20

 Ordering cost S = $300

 Holding cost h = $1 per unit and week
Example 3: Part-period balancing heuristic

 If the delivery in period 1 covers only the demand in

period 1, the total holding costs are 0
 0 < 300 (Diff = 300)
Example 3: Part-period balancing heuristic

 We now successively consider the total holding costs for

deliveries that cover 2, 3,. . . periods
 2 periods 60  300 (Diff = 240)

 3 periods 60 + 2  90 = 240  300 (Diff = 60)

 4 periods 60 + 2  90 + 3  70 = 450  300 (Diff = 150)

 A new delivery in period 4

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 3: Part-period balancing heuristic

 The same procedure is applied with period 4 as the first

period
 2 periods 30  300 (Diff = 270)

 3 periods 30 + 2  100 = 230  300 (Diff = 70)

 4 periods 30 + 2  100 + 3  60 = 410  300 (Diff = 110)

 A new delivery in period 7

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 3: Part-period balancing heuristic

 The same procedure is applied with period 7 as the first

period
 2 periods 40  300 (Diff = 260)

 3 periods 40 + 2  80 = 200  300 (Diff = 100)

 4 periods 40 + 2  80 + 3  20 = 260  300 (Diff = 40)

d1 d2 d3 d4 d5 d6 d7 d8 d9 d10
50 60 90 70 30 100 60 40 80 20
Example 3: Part-period balancing heuristic

 The solution
 A delivery in period 1
 A new delivery in period 4

 A new delivery in period 7

Period t 1 2 3 4 5 6 7 8 9 10
dt 50 60 90 70 30 100 60 40 80 20
Solution 200 200 200
Example 3: Part-period balancing heuristic

 Total cost of the solution

 Ordering cost = 3  300 = 900
 Holding cost = 60 + (2  90) + 30 + (2  100) + 40 + (2  80)

+ (3  20) = 730
 Total cost = 1630

 5% higher than the optimal costs obtained by the Wagner-

Whitin algorithm
Period t 1 2 3 4 5 6 7 8 9 10
dt 50 60 90 70 30 100 60 40 80 20
Solution 200 200 200
 Example: Summary of the results

Deviation from the

Method Total cost ($)
optimal solution (%)
The Wagner-Whitin algorithm 1550 0.00
The Silver-Meal heuristic 1630 5.16
Least unit cost heuristic 1870 20.65
Part-period balancing heuristic 1630 5.16
Performance of the heuristics

 Performance of the heuristics on 20 test problems

(Baker, 1989)

Method Number optimal Average percent error

The Silver-Meal heuristic 14 0.943
Least unit cost heuristic 8 9.303
Part-period balancing heuristic 16 1.339
 Example: Heuristic methods

 MRP record for the Silver-Meal heuristic

Lot Size: Silver-Meal Lead Time: 2 weeks

Item: A
Weeks
1 2 3 4 5 6 7 8 9 10 11 12
Gross requirements 130 70 90 70 30 100 60 40 80 20
Net requirements 50 60 90 70 30 100 60 40 80 20
Scheduled receipts 10
Projected on hand
80 80 80 150 90 0 30 0 100 40 0 20 0
inventory
Planned receipts 200 100 200 100
Planned order
200 100 200 100
releases
 Example: Heuristic methods

 MRP record for part-period balancing heuristic

Lot Size: Part-period balancing Lead Time: 2 weeks

Item: A
Weeks
1 2 3 4 5 6 7 8 9 10 11 12
Gross requirements 130 70 90 70 30 100 60 40 80 20
Net requirements 50 60 90 70 30 100 60 40 80 20
Scheduled receipts 10
Projected on hand
80 80 80 150 90 0 130 100 0 140 100 20 0
inventory
Planned receipts 200 200 200
Planned order
200 200 200
releases