This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: University 0f Southern California
Depuﬂmm‘ of Industrial and Systems Engineering
33!? jfﬂf. f’rotfm‘fiﬂn I: Facifiﬂ'es and Logistics
Spring 2600
April Handouts {PM 2)
Computerized Relative Allocation of rhcﬁ'iﬂﬂ Technique (CRA FT)
Dr. Ardm'tm Ase} Vazfri We explained that facilities layout algoridtins are divided into construction algorithm and improvement algorithms. Construction algorithms start from scratch and end up 1with a layout.
In handout I5 we explained a construction algorithm. Now we explain an improvement algorithm
called CRAFT {Computerized Relatith: Allocation of Facilities Technique}. CW starts from
an initiaJ layout and end up with a better layout. We also explain MicroCRAJT and MULTIPLE
which are slight modiﬁcation or'CRAFl'. (liten an initial layout. CRAlT interchanges pairs or triples of departments. We onlyr discuss
pairwise interchange. Therefore, it picks Up a pair of departments and changes their locations. The objective is to reduce total flow it distance. Distances are rectilinear and centroid to centtoid. Given the Following layout and FTchart. ED 30 4U 40
First we idemin the centroid of the department and measure the rectiiirtear distance. 5] 70 . gain}. Total ﬂow it distance in this layout is 1020. Can we by pairwise interchange of department
reduce this total ﬂow x distance? CRAFT only changes the locations of a pair of departments which are either adjacent or of the
same size. Why? CRAFT interchanges the location of all potential pairs of departments, and ﬁnd the pair of
departments which their interchange results in maximum saving in ﬂow x distance. CRAFT assumes when two departments are interchanges their centroids are interchanged. This
is true for two departments of the same size but not for two adjacent departments. Therefore,
CRAFT indeed estimates the total saving resulted from interchange of two departments. Suppose CRAFT interchanges the locations of A and B. Since it assumes that when A and B are
interchanged, their centroids are interchanged. Therefore, the new distance from departmentA to any other department is equal to the old distance of department B to that department. Distances
AB unchanged A—C (old BC) = 65
AD (old BD) = 25 (D) (F) (FXD) Estimated (F)(D) is 1060, which means interchange of A and B does not result in a saving. Now let’s check interchange of B and D. Let’s also do not include unnecessary calculations. The distance matrix is as follows. Net result is 75. Therefore the estimated saving is 75.
After examining all other potential interchanges. A—C, AD, C—D we conclude that interchange of
B and D results in maximum saving. Note that this is estimated saving, not actual saving, since we have assumed that when B and D are interchanged, their centroras are interchanged. Then we interchange B and D and calculate the actual change in ﬂow x distance. A
. (2530)
C B
. (20,10) . (55,10) D1 . (55,30)
D2
0 ) Area ofDI X Area ofD2 XD
‘ Area of (Di+D2) D1 + Area of (D1+D2) 2 @ a _
XD= 800 65 + 80075 — 67.5 $529 w
YD: 800 30* 800 w 25 New actual distance matrix is Total ﬂow x distance in the new layout is 985. Flow x distance in the initial layout was 1020.
Therefore interchange of B and D results in 1020 — 985 = 35 units saving. Note that estimated saving was 75. Now we look at this layout as the initial layout, and the interchange process re—starts from the
beginning. The candidates for interchange are AB, AC, AD, BC, CD. Estimated savings are calculated and A—B shows the highest estimated saving. Again the above layout is assumed as the initial layout. Potential interchanges are AC, AD, B C, C—D. CD results in maximum estimated saving. The next layout is Again the potential candidates for interchange are AB, AC, AD, BD. None of them will result in a ﬂow x distance saving. The process stops. Is this layout optimal? No, it is Z—optimal. Can we ﬁx some departments in CRAFT?
Yes. ...
View
Full
Document
This note was uploaded on 02/01/2009 for the course ISE 310L taught by Professor Bottlik during the Spring '06 term at USC.
 Spring '06
 Bottlik

Click to edit the document details