1 methodology for monitoring supply chain performance: a fuzzy logic approach source : logistics...
TRANSCRIPT
1
Methodology for monitoring supply chain performance:
a fuzzy logic approach
Source : Logistics Information Management
Volume 15. Number 4. 2002. pp. 271-280
Author : H.C.W Lau , Wan Kai Pang and
Christina W.Y. Wong
Speaker : 曾偉育Member : s9114638 曾偉育 s9114624 王仁群
2
Outline
IntroductionAdoption of fuzzy logicDefuzzificationCase exampleConclusion
3
Introduction
In this paper, fuzzy logic principles is recommended to monitor the supply chain performance by evaluating the ongoing delivery time and product quality,and performing adjustment in order quantity based on the performance.
4
Planned Performance
Standards
ActualPerformance
Monitor and Compare
AdjustmentAnd
Investigation
inputinput
input
Supplier monitoring system
5
Cost function for supply chain management
Three kinds of costs for supply chain management.
Total cost = variable cost
+ fixed costs
+unprecedented costs.
The unprecedented costs are difficult to measure using the traditional quantitative approach.
6
Delivery time Standard
More than 4 days early(<4) Fail
4 days early(4) Acceptable
3 days early(3) Acceptable
2 days early(2) Acceptable
1 days early(1) Fine
Promised delivery day Fine
1 days late (-1) Fine
2 days late (-2) Acceptable
More than 2 days late(>-2) Fail
Delivery time measurement
7-2
0.1
0.5
1.0
-4 -1-3 0 1 2 3 4Days
Degree of Membership
Adoption of fuzzy logic
8
Quality value(defect percentage) Standard<1 Fine
2 Acceptable
3 Acceptable
4 Acceptable
5 Acceptable>5 Fail
Quality measurement
91 2 3 4
1.0
0.5
5
0.45
Degree of Membership
Quality(percentage)
Adoption of fuzzy logic
10
Recordscomplated Weighted method
1 Delivery time of Record1*100%
2 Delivery time of Record1*75%+delivery time of Record2*25%
3 Delivery time of Record1*50%+delivery time of Record2*30%+delivery time of Record3*20%
4 Delivery time of Record1*40%+delivery time of Record2*30%+delivery time of Record3*20%+delivery time of Record4*10%
Weight average for supplier defect rate assessment
11
Recordscomplated Weighted method
1 Defect percentage of Record1*100%
2 Defect percentage of Record1*75%+defect percentage of Record2*25%
3 Defect percentage of Record1*50%+defect percentage of Record2*30%+defect percentage of Record3*20%
4 Defect percentage of Record1*40%+defect percentage of Record2*30%+defect percentage of Record3*20%+defect percentage of Record4*10%
Weight average for supplier defect rate assessment
12
Delivery time Quality Quantity
Fail Fail Substantial decrease (SD)
Fail Acceptable Considerable decrease (CD)
Fail Fine Some decrease (SMD)
Acceptable Fail Considerable decrease (CD)
Acceptable Acceptable Some decrease (SMD)
Acceptable Fine Little decrease (LD)
Fail Fail Some decrease (SMD)
Fail Acceptable Little decrease (LD)
Fail Fine No change (NC)
Change of next order quantity
13
Defuzzification
Defuzzification is the process of reducing a fuzzy set to a single point.
There are several methods of performing defuzzification, the gravity method is the most common one.
Next order quantity =
quantity +(average quantity) × (order quantity change rate)
The output of Defuzzification
14
No. Defect rate(%) Delivery time (days)
1 5 -2
2 3 -2
3 2 -1
4 0 -2
Recommend the ordering quantity of next order :
weighted average for defect rate=
5%*0.4+3%*0.3+2%*0.2+0%*0.1=3.3%
weighted average of delivery time=
(-2)*0.4+(-2)*0.3+(-1)*0.2+(-2)*0.1= -1.8
Case example
151 2 3 4
1.0
0.5
5
3.3%
0.45
Degree of Membership
Quality(percentage)
Adoption of fuzzy logic
weighted average for defect rate= 5%*0.4+3%*0.3+2%*0.2+0%*0.1=3.3%
16-2
0.1
0.5
1.0
-4 -1-3 0 1 2 3 4Days
Degree of Membership
-1.8
Adoption of fuzzy logicweighted average of delivery time=
(-2)*0.4+(-2)*0.3+(-1)*0.2+(-2)*0.1= -1.8
17
Delivery time Quality Quantity
Fail Fail Substantial decrease (SD)
Fail Acceptable Considerable decrease (CD)
Fail Fine Some decrease (SMD)
Acceptable Fail Considerable decrease (CD)
Acceptable Acceptable Some decrease (SMD)
Acceptable Fine Little decrease (LD)
Fail Fail Some decrease (SMD)
Fail Acceptable Little decrease (LD)
Fail Fine No change (NC)
Change of next order quantity
18
Fuzzy pattern of order quantity change rate
0-0.6 -0.5 -0.4 -0.3 -0.2 -0.1
0.1
0.2
0.3
0.6
0.5
0.8
0.7
0.4
0
0.9
1.0SD CD SMD LD NC
SDSubstantial Decrease
CDConsiderable Decrease
SMDSome Decrease
LDLittle Decrease
NCNo Change
Degree of Membership
Order Quantity Change Rate
COG=-0.225
19
Next order quantity
Next order quantity = quantity +(average quantity)(order quantity change rate)
Average order quantity = 10000 unitsQuantity needed for the coming production is 20000 units
Next order quantity = 20000 + 10000(-0.225) = 17750 units
Company will complete the order by choosing another supplier to order the rest of the 2250 units to complete the order
20
ConclusionThe methodology of using fuzzy logic in monitoring a supply chain partners’s performance and provide a suggestion on the next order quantity.
Developing fuzzy rules require experience from field expert, experimental results and theoretical derivation.
21
thanks