8 views

Uploaded by InesGalic

fulltext 14.pdf

- bitszg553mid5
- Operating_IIUC
- Blueprint Training
- CPU_Sched
- Logica Fuzzy
- Available to promise ATP
- A Historical Review of Evolutionary Learning Methods for Mamdani-type Fuzzy Rule-based Systems_ Designing Interpretable Genetic Fuzzy Systems
- Recent Neutrosophic Models for KRP Systems
- Simulation of Scheduling Algorithms
- Release Time
- SMAW Process Parameters Optimization using Taguchi & Fuzzy Logic
- A framework for PEST analysis based on neutrosophic cognitive map: case study in a vertical farming initiative
- Operating System
- QT
- 00816520
- IJGCA 030405
- Thread Scheduling and Dispatching
- pizzz
- Introduction to Fuzzy Sets
- cpu_sched.

You are on page 1of 12

DOI 10.1007/s10845-008-0093-5

in a robot-centered flexible manufacturing cell

Indira Molina Restrepo S. Balakrishnan

Received: 30 January 2008 / Accepted: 1 February 2008 / Published online: 28 February 2008

Springer Science+Business Media, LLC 2008

the sequence of part movements in a multi-product batch processing through a computerized machine cell is presented in

this paper. A number of production objectives are taken into

account. Two fuzzy based strategies: fuzzy-job and fuzzymachine are proposed and their performance is compared

to two well known dispatching rules such as SPT (Shortest Processing Time) and WEED (Weighted Earliest Due

Date). The sequencing algorithm was implemented on a standard personnel computer and the scheduler was interfaced to

a robot controller for implementing loading and unloading

strategy within the cell. The proposed fuzzy-based methodologies especially fuzzy-job shows a superior performance

compared to the traditional dispatching rules considered.

Keywords Fuzzy logic Multi-objective scheduling

Flexible manufacturing cell

Introduction

Flexible Manufacturing Cell (FMC) is a group of automated

machines serviced by a central material-handling device such

as robots or automated guided vehicles. The flexibility to

process a wide variety of parts gives rise to numerous ways

of routing a group of parts within the cell. One of the most

important problem encountered in the FMC scheduling deals

with the assignment of given resources to different processes.

The best schedule must satisfy the production objectives

considered. The production objectives taken into account

in this study are: maximize throughput; minimize penalty

I. M. Restrepo S. Balakrishnan (B)

Department of Mechanical and Manufacturing Engineering,

University of Manitoba, Winnipeg, MB, Canada R3T 2N2

e-mail: balakri@cc.umanitoba.ca

robot travel time. The processing environment considered is

a multistage-job shop problem wherein each job has its own

processing requirement and sequences. The scheduling problem considered is an open, deterministic, and static case. The

scheduling criteria include both cost and performance related measures. The key objective of this work is to develop

an efficient computer based scheduling technique to generate

the sequence of parts, and control the movement of the robot

within a manufacturing cell in order to fulfill the chosen production objectives. Although standard industrial robots are

computerized, they do not possess the intelligence to dynamically alter preprogrammed motions. Hence an external personal computer (PC) is employed to provide that additional

control capability.

In the present scenario, each batch consists of multiple

orders. Thus, each batch involves multiple products, each

with different specifications. The machining times vary from

part to part. Certain parts may need to be processed on multiple machines, and a certain sequence of operations may

be important. In a sequential process, the parts are routed

through the cell to visit machines in a pre-specified order.

Customer requirements refer to due dates and penalties,

whereas production costs refer to robot and machine idle

costs. On the other hand, a non-sequential processing will

allow the flexibility to route the parts bypassing the sequencing requirement. Penalty values are assigned to each part type

and are depended on the characteristics of the customer and

terms of contracts. Robots and machines are an expensive

investment and hence it is necessary to use them effectively.

It is important to balance the idle times of the robot versus

that of the machines. Machine idle cost requires a production

study, and these values have to be set by the user. In our setting, all data is randomly assigned in order to mimic a wide

variety of manufacturing situations.

123

422

on a machine. Processing times are part specific and are not

machine dependent. Parts are delivered to the machines when

needed, and hence no buffers are present in each machine.

The task of the robot is to pick up parts from an input buffer,

move them through the cell, and drop them off at an output buffer. The travel time of the robot between machines is

known and remains fixed. However it should be noted that the

total travel time taken by the robot to move a part through its

production cycle will be a function of the sequence generated

by the scheduling algorithm. The multi-objective nature of

the problem under consideration presents many challenges.

Many approaches can be followed. For this project, a new

fuzzy-based approach with a limited number of heuristic

rules is considered. It was found to be easy to model and

the computational times were found to be very minimal, an

aspect that is very critical for real-time implementations.

Review of literature

Extensive research has been done on aspects related to scheduling of flexible manufacturing systems. Due to the complexity and nature of scheduling, several approaches have

been considered, ranging from traditional solutions involving complex mathematical analysis to the recent approach,

namely reasoning algorithms. In addition, a number of different dispatching rules have also been proposed. An efficient

use of robotic cells will no doubt increase the production

rates. Modeling issues related to loading of multiple machines

with a dual-gripper robot has been proposed (Sethi et al.

2004). Analysis for finding optimal moves of material handling systems in a production cell for the manufacture of

machine castings for truck differential assemblies has been

researched (Brauner and Finke 2001). The findings from this

study are applicable to sequential flow of products through

manufacturing cell and cannot be applied to products with

varying sequences and process times. They also consider only

a single objective. It has been shown that time based rules

such as SPT are good to reduce flow time and machine idle

time (Chen and Lin 1999). However, jobs with long processing times tend to be tardy. Dispatching rules based on

tardiness, the bottleneck dynamics, resource pricing and the

effects of inserted idleness have been proposed (Kutanoglu

and Sabuncuoglu 1999). They concluded that different pricing schemes should be used in different environments. A

coordination rule that looks ahead about machine idle time,

and mechanisms for tracking the demand, and selection of

jobs has been proposed (Holthaus and Ziegler 1997). The

simulation studies demonstrated the effectiveness of the rule

in improving flow-time and duedate based objectives compared to scheduling rules without any coordination.

A concurrent solution for loading and scheduling in a

flexible manufacturing environment has been found to be

123

technique to optimize the moves of a robot for loading

parts in a three machine cell has also been developed

(Chen et al. 1997). Their solution became ineffective when

the number of parts exceeded ten. A solution based on colored

Petri Nets to control the alternative machining and sequencing in a FMC has been presented (Yalcin and Boucher 1999).

The study showed that they could be used as an effective

modeling tool.

Techniques based on artificial intelligence have been explored by a number of investigators. A methodology based

on training a back-error propagation network, to solve a jobshop scheduling problem has also been investigated (Jain and

Meeran 1996). However, their methodology is not effective

for large scale problems. Fuzzy set theory has been found

to be useful for modeling and solving scheduling problems

with uncertain data. A multi-objective fuzzy approach that

employed membership functions to find the share of each

objective in arriving at the final decision rules has been found

to be effective (Kazerooni et al. 1997). It was then applied

for selecting machines, after pre-selection of a job using traditional scheduling rules. Their approach combined different

rules for the selection of jobs and machines. A methodology

based on a combination of fuzzy and SPT strategy improved

the net profit and average lead-time in the test cases they

considered.

In a robot centered machine cell, the robot may be programmed to wait in front of a machine and pick up the part as

soon as the machine completes processing a part and move

it to the next processing station. This situation is termed as

no-wait constraint. A two and three machine problem for

processing identical and dissimilar parts have been proposed

and results pertaining to optimal cycle time for this type of

situation have been presented (Agentis 2000) and (Agentis

and Pacciarelli 2000). A neural network based approach for a

single machine job sequencing has been found to be effective

(El-Bouri et al. 2000). A problem is classified first by one type

of neural network into one of a number of categories. Then

another neural network, which is specialized for a particular

category, applies previously learnt relationship to produce a

job sequence that aims to better satisfy the given objective.

Use of genetic algorithms (GA) for tackling scheduling

problems has also received wide attention. An approach to

optimize the operation when products have to meet a due date

constraint has been the focus in one such study (Sannomiya

and Iima 1996). GA has also been used to obtain solutions

for scheduling a FMS with setup time as constraint (Jawahar

et al. 1998). A fuzzy-based methodology to address the

machine loading problem has been developed (Vidyarthi and

Tiwari 2001). Even though the minimization of system unbalance and maximization of throughput were their objective,

they did not take into account the inclusion of due-date related

objectives to reduce costs. The job ordering and sequencing

made on the evaluation of membership functions.

From a review of literature, it can be concluded that no

generalization can be made regarding various dispatching

rules. No single rule has shown to be superior in all type

of scenarios. They have not performed uniformly well on

more than one criterion. A common belief is that a combination of simple dispatching rules or a combination of heuristics with simple dispatching rules performs better than single

rules in many cases. When multiple part types with each having its own sequence and processing times is put through

a manufacturing cell, in addition to the traditional process

objectives, the efficient use of material handling system also

becomes critical. Although the travel times of robots between

machines are fixed, the total travel time required to move a

part through the cell will be a function of the sequence generated and cannot be assumed to be the same always.

The work reported in this paper uses fuzzy logic because

of its ability to deal with uncertain data. As will be shown,

it is easy to implement and is an excellent tool for decision

making in multi-objective systems. It needs fewer rules, and

the knowledge is easier to model. It also leads to faster solutions, a key issue for real-time control of FMC.

Methodology

The proposed fuzzy-logic based methodology is an extension

of SPT and WEDD, and has the ability to take into account

factors such as machine and robot idle costs. When a group

of parts are processed on a FMC, the modeling issues when

multiple objectives are considered are extremely complex.

The routing sequence and processing time directly contribute to the idle cost of machines and robot. Different objective

function will give rise to different routing sequence. Fuzzyset approach in combination with a set of heuristic rules was

chosen as the methodology with a hope to reduce the computational burden on the control processors. This issue is

critical for real-time control of robots and machines. The

fuzzy approach employed, considers part and machine characteristics as two separate fuzzy subsets. SPT and WEDD are

well known for their effectiveness in optimizing the throughput and tardiness cost, respectively. However, they are not

effective in dealing with multi-objective problems. Before

describing the fuzzy-based approach, a brief review of the

scheduling algorithms developed and the ones compared with

will be presented next.

Shortest processing time (SPT) and Weighted earliest due

date (WEDD)

Shortest processing time (SPT)jobs are sequenced in

increasing order of their processing time from many combinations of job/machine. As is well known, SPT provides a

423

long processing times and early due dates tend to be tardy.

SPT is defined by:

SPT = Min P T ( j, m),

(1)

Weighted earliest due date (WEDD)jobs are sequenced

in increasing order of the ratio of the due date of the job to

the penalty of the job assigned when the job is late. Although

WEDD performs well in terms of due date and penalty, it

does not take into account processing time information, and

thus poor use of resources. WEDD can be expressed as:

DD

,

(2)

WEDD = Min

P

where Min(DD/P) = minimum ratio of the due date (DD) over

the penalty (P) of the jobs.

Fuzzy-logic

The foundation of fuzzy logic is fuzzy theory (Zadeh 1965).

A fuzzy set theory to deal with problems in which the source

of imprecision is the absence of sharply defined criteria was

formulated. A fuzzy set, A, in X is characterized by a membership function A (x)which associates each element in X,

a real number in the interval [0,1]. The value of A (x) at x

represents the grade of membership of x in A. The closer this

value is to unity, the higher is the level of membership of x in

A. On the contrary, in crisp logic the membership values are

either 0 or 1. Fuzzy logic introduces the possibility of intermediate values, and a more precise way of defining grades

of membership of an element in a domain.

Fuzzy set operators can be defined in terms of operations

between membership functions. These operations are important because they can describe interactions between variables. The basic operations in fuzzy logic are intersection,

union and complement (Zadeh 1965) and are defined as follows:

Intersection :

(3)

Union :

(4)

x X : A (x) = 1 A (x).

(5)

Complement :

Other extensions using simple algebraic transformations have

also been provided (Cox 1998).

The concept of fuzzy decision making (Bellman and Zadeh

1970) can be defined as: a decision process in which the

goals and/or constraints, but not necessarily the system under

control, are fuzzy in nature. They pointed out that fuzzy

goals and constraints can be defined accurately as fuzzy sets.

123

424

given goals and constraints. Based on this work and further

development (Yager 1978), the decision-making technique

developed in the present work will incorporate material presented below.

Objectives (goals and constraints) can be easily represented by fuzzy sets. Assuming we have a set of alternatives

in a decision X = [X 1 , X 2 , . . ., X n ] and a particular objective A, we can associate with each element in X i a number

A (X i ) in the interval [0,1] indicative of how well X i satisfies

objective A. The advantage of fuzzy sets derives from the fact

that very fuzzy objectives as well as very precise objectives

can be represented. In order to extend the above definition to

combine multi-objectives in decision making, let us assume

we would like to select among the set of alternatives X the

one that best satisfies a set of objectives A1 , A2 , . . ., A p . Each

alternative X i , is assigned a number indicative of how well

it satisfies the objectives as a group A (X i ) and, of course,

the X with the highest value is the best. The dilemma would

be on how to combine the contribution of each element to

each objective, in order to get an overall general contribution of each element to all the objectives as a whole. Using

an intersection operation one would be able to combine the

objectives.

The values of the alternatives are obtained by using membership functions. There are different ways of representing

the intersection of objectives. One method (Zadeh 1965)

chooses the minimal value among the objectives. However, if

an alternative does not contribute to an objective at all (i.e.,

A (X i ) = 0), then the result of the intersection would be

zero (0), thus excluding the contribution of the alternative to

the other objectives even if those are close to unity. Similarly,

product and bounded sum methods eliminate the contribution of the objectives if one among them has membership

value of zero (0). On the contrary, mean method produces

an average value of the contributions of the alternative to the

different objectives. Hence, if the alternative does not benefit

an objective at all, the solution does not go to zero (0). Furthermore, in our specific case we are not only interested in

the best alternative, but also in the second or third best one.

For example, in the case of an alternative (part type), which

cannot be chosen as a result of resources not being available

(i.e., machine that needs to process is busy or part being blocked, etc), then second or even third best alternative has to be

looked at. Table 1 presents a batch composed of three different part types: A, B and C; and three objectives represented

by membership functions A ( p), B ( p) and C ( p). In this

table, columns 1 and 2 refer to the part type and the number of parts to be produced respectively. Column 3 refers to

the job number. Columns 4, 5 and 6 represent the processing

times for the respective machines, namely M1, M2 and M3.

Table 1 also shows the numerical results obtained by using

the four methods (Restrepo and Balakrishnan 2001), namely

123

observed, the mean method is the only one that can assure us

that we would have more than one alternative to choose from.

For example, if part type B cannot be selected for any reason

by using any method but mean, we would not have a second choice. In fact, mean would leave us to choose a second

alternative (part type C) and a third alternative (part type A).

In the present study, the evaluation of the overall contribution of the fuzzy membership function of each part type

determines the sequence of the jobs in a given batch. Two

methods: fuzzy-job and fuzzy-machine are proposed. These

strategies are mainly used to choose the jobs and machines

in the sequence.

Fuzzy-job

Job sequencing is determined by evaluating the overall contribution of the fuzzy membership function of the part type

to the optimal performance of the system. The various membership functions included for fuzzy-job are defined below.

Th ( p): The throughput of part type p is defined by a

membership function that is defined as the ratio of the difference between the maximum total processing time PT of

the part types, and the total processing time of the part type

p to the difference between the maximum and minimum

total processing time of the part types. This membership

evaluates the contribution of the part type to maximize the

throughput of the batch. This can be expressed as:

Th ( p) =

MaxPT PT (p)

, 0 Th ( p) 1,

MaxPT MinPT

(6)

MinPT= minimum processing time of part types, and PT(p)

= total processing time of part type p.

P T ( p) is defined byP T ( p) = N ( p)

M

P T ( p, m),

m=1

(7)

where N(p) = number of jobs for part type p, m = machine

number, m = 1,2,3,,M, M = number of machines, and

PT (p,m) = processing time of part type p on machine m.

P ( p): The membership function for the penalty of part

type p is defined as the ratio of the difference between the

maximum total penalty TP of the part types, and the total

penalty of the part type p to the difference between the

maximum and minimum total penalty of the part types.

This membership evaluates the contribution of the part

type to minimize the total penalty due to late jobs. This

can be expressed as:

425

Part type

A

B

C

# Of parts

Jobs #

M1

M2

M3

A (p)

B (p)

C (p)

2

3

1

1,2

3,4,5

6

4

0

0

2

3

5

0

0

1

0

0.50

1

0.40

1

0

0.50

1

0.5

0

0.50

0

0.32

0.83

0.50

0

0.50

0

0

0.50

0

M1, M2, and M3 are machine 1 through 3; ZZadeh; MMean; PProduct; BBounded sum

P ( p) =

MaxTP TP(p)

, 0 P ( p) 1,

MaxTP MinTP

MinTP = minimum total penalty of part types, and TP(p) =

total penalty of part type p.

TP(p) is defined byTP(p) =

DD(p)

,

P(p)

(9)

of part type p.

IC ( p): The membership function for the machine idle

cost of part type pis defined as the ratio of the difference

between the total machine idle cost of part p, and the minimum total machine idle cost of the parts to the difference

between the maximum and minimum total machine idle

cost of the part types. This membership evaluates the contribution of the part type to minimize the total machine

idle cost when producing the batch. This can be expressed

as:

IC ( p) =

C(p) MinC

, 0 IC ( p) 1,

MaxC MinC

(10)

types, MinC = minimum total machine idle cost of part

types, and C(p) = total machine idle cost for part type p.

C( p)is defined byC(p) =

M

U I C(m),

T ( p) =

(8)

(11)

m=1

where

U =

0,

if

machine is not needed for processing

IC(m) = machine idle cost of machine m.

of part type pis defined as:

1,

1/M,

if job visits more than 1 machine,

This membership evaluates the contribution of the part

type to minimize the total robot travel time when producing the batch.

p

O : The overall membership function of part type p is the

average mean of the individual membership function of

the penalty, throughput, machine idle cost and robot travel

times for part type p. By using the mean method defined

earlier, this can be expressed as:

P ( p) + Th ( p) + IC ( p) + T ( p)

,

4

0 O ( p) 1.

O ( p) =

(12)

Fuzzy-machine

The approach followed is similar to fuzzy-job. However, instead of evaluating the contribution of the part type to meeting

the objectives of the system, the evaluation is given by obtaining a membership function for each combination of part type

and machine. Thus, for fuzzy-job, there would be as many

membership functions as number of part types, while for

fuzzy-machine, there would be as many membership functions as number of part types and machines each part has to

visit. For example, for the batch given in Table 2, the number of membership functions for fuzzy-job would be three:

(0 (A), 0 (B), 0 (C) ), and for fuzzy-machine would be six:

(0 (A,M1), 0 (A,M2), 0 (B,M2) ), (0 (B,M3), 0 (C,M3),

0 (C,M4) ). The variables employed in Table 2 are those used

in Table 1.

Although the robot travel time between machines is fixed

and known, the total robot travel time required to pick up a

Part type

A

B

C

# Of parts

Jobs #

M1

M2

M3

M4

2

2

2

1,2

3,4

5,6

4

0

0

2

13

0

0

7

5

0

0

8

123

426

Part type

Batch #1

A

B

C

Batch #2

A

B

C

Batch #3

A

B

C

# Of parts

Job #

M1

M2

M3

M4

2

2

2

1,2

3,4

5,6

3

2

2

5

5

5

DD(p)

P(p)

PT(p)

TP(p)

C(p)

4

0

0

2

13

0

0

7

5

0

0

8

40

35

55

3

2

4

12

40

26

13.3

17.5

13.75

3

4

5

1,2,3

4,5

6,7

45

0

44

48

43

0

0

0

46

0

42

0

280

220

220

4

2

3

279

170

180

70

110

73.3

3

4

4

1,2,3,4,5

6,7,8,9,10

11,12,13, 14,15

28

0

0

32

25

0

0

30

23

0

0

35

480

460

440

3

2

2

300

275

290

60

230

220

3

4

5

DD(p)Due date of part type p

P(p)Penalty of part type p

PT(p)Total processing time of part type p

TP(p)Total penalty of part type p, defined by: DD(p)/P(p)

C(p)Total machine idle cost for part type p

output buffer will depend on the sequence and scheduling

cycle for the part. This is unknown until the best schedule

is arrived at. Hence a membership function is included in

the analysis. Values of penalty time remain constant for each

part type. These values do not vary according to the machine.

Therefore, these membership functions are defined as in

fuzzy-job. The robot travel time between machines is assumed to be the same as in fuzzy-job. Throughput and machine

idle cost memberships are defined somewhat differently as

described below.

Th ( p, m): The membership function for the throughput

of part type pon machinemis defined in the same terms as

in fuzzy. The difference is that the processing time is given

by PT(p,m) instead of PT(p). Th ( p, m) can be defined as:

MaxPT PT (p, m)

,

MaxPT MinPT

0 Th ( p, m) 1,

Th ( p, m) =

(13)

where

MaxPT = maximum processing time of combinations parttype/machine, and MinPT = minimum processing time of

combinations part-type/machine.

IC ( p, m) : The membership function for the machine

idle cost of part type p on machine m is defined as before,

except that instead of the total machine idle cost C of part

p, the membership function is defined in terms of machine

idle cost of part p on machine m. IC ( p, m) can be defined

as:

123

IC(m) MinIC

,

MaxIC MinIC

0 IC ( p, m) 1,

IC ( p, m) =

(14)

and MinIC = minimum machine idle cost of machines.

An illustrative numerical example is provided below for a

sample data (batch #1) shown in Table 3. For fuzzy-job, there

are three membership functions that are evaluated:

o (A), o (B), o (C). The numerical results resulting membership functions are shown in Table 4. As can be noticed,

part type C (machine M3) would be the first in the sequence

followed by part type A (machine M1), and finally part type B

(machine M2). For the fuzzy-machine method, membership

functions are expressed in terms of part type and machine.

For the combination of part type A and machine M1, the

various membership functions are given in Table 5.

The results presented in Table 5 indicate that the sequence

would be Part type C-Machine M3, Part type A-M1, Part type

A-M2, Part type C-M4, Part type B-M3, and Part type B-M2.

Validation of the proposed methodologies

In order to compare the performance of various scheduling options, a set of data that produces a wide variety of

Part type

P ( p)

Th ( p)

IC ( p)

T ( p)

O ( p)

A

B

C

1

0

0.89

1

0

0.50

0

0.50

1

0.50

0.50

0.50

0.62

0.25

0.72

427

Part type/

machine

P ( p, m) Th ( p, m) IC ( p, m) T ( p, m) O ( p, m)

Part A, M1

Part A, M2

Part B, M2

Part B, M3

Part C, M3

Part C, M4

1

1

0

0

0.89

0.89

0.81

1

0

0.54

0.72

0.45

0.50

0

0

1

1

0.50

0.50

0.50

0.50

0.50

0.50

0.50

0.70

0.62

0.12

0.51

0.77

0.58

The effect of due date penalties is reflected by variations in

due date indicated. All the data were randomly generated to

study the effect of variations in processing times, processing requirements and due dates. Comparisons were made in

terms of machine idle cost, tardiness cost, and throughput

resulting from employing different strategies.

The different models presented were implemented on a

personal computer using Borland C++. Calculations were

made to determine the contribution of each part type to the

chosen objectives. Upon initiation of the program, the user

is prompted to enter the customer specified data (1 in Fig. 1)

followed by processing data (2 in Fig. 1). The program then

provides a choice to initiate sequential (3 in Fig. 1) or nonsequential processing (4 in Fig. 1). For both the options one

may choose either prioritizing Loading or Unloading

(5 or 6 in Fig. 1).

If sequential processing is chosen, a part is loaded only if

the first machine in the series of machines that the part needs

to go is free. It is likely that there may be more than one part

as a suitable candidate for loading and the following order of

priority is implemented. First priority goes to the first part in

the order of feasible parts that needs the currently least occupied station (e.g., a station having the fewest waiting parts).

If there are more than one station, then the priority will be

assigned to the station that is most in demand by the parts in

the current batch. Algorithm will search for all the possible

candidate part and generate the sequence as well as setup the

robot programs. For non-sequential processing, the process

of part selection goes through the same order of priorities as

defined for sequential process. However, in this case, a part

is a candidate providing one of its machining requirements

is met. There are two additional criteria in the case of nonsequential processing. The part to be loaded, after having met

all the other requirements, must go through a conflict check.

The conflict check is a subroutine that ensures that, if a part

is loaded into the system, it will not cause a blockage to the

further processing of parts.

The next choice is either to prioritize Loading of

Unloading. For loading priority, the following options in

decreasing order of importance are followed: (i) load a

machine with a part from the input buffer; (ii) shift part from

buffer; and (iv) move the robot and wait at the next station

needing unloading. The unload priority follows the options

listed next in decreasing order of priority: (i) Unload a part

; (ii) If option is not possible, then consider shifting a part

from one station to the next; (iii) if shifting is not an option,

then it may load a part: and (iii) if none of the above is possible then move the robot to a station that has almost finished

machining and will be waiting to be unloaded.

The last step in the program (7 in Fig. 1) provides the user

with the ability to run the different options of scheduling

algorithms. The various options and the key steps are shown

in figure. When the program is executed, it enters into a loop

and continues until all the parts are processed. Inside each

loop, an internal variable checks for both a loading or unloading priority. This internal variable will lead the sequence of

operations within the loop: loading, shifting, unloading, and

moving and waiting. A look-ahead feature built within the

program, seeks to reduce the throughput time of a batch by

saving extra travel time of the robot. Instead of waiting for the

next part to be finished, the robot keeps track of which parts

are almost done and goes next to the corresponding station,

thus saving movement time. If the user wishes to compare

various options of scheduling and automatically pick the optimal scheduler, then the choice ALL will be selected and the

software will automatically generate the results shown later

in both graphical form as well as text file. The output from

this choice is eventually used for controlling the entire operation of scheduling and controlling robot moves.

The sequence follows the order in which the machines

are numbered. However, it is important to clarify that the

methodology proposed has the capability of processing any

sequence. The sequence followed is only for illustrative purposes. For a sample data (batch #2) presented in Table 3, the

sequence for part type A is M1M2, for part type B, it is

M2M4 and M1M3 for part type C. The part processing

is simulated using a loading priority. Some attributes can be

noticed for this batch. The part type with the longest processing time (part type A) is also the part with the longest due

date, and it has the highest penalty for tardiness. Besides, the

part with the shortest processing time and shortest machining

time (part type Bmachine M2) has the lowest penalty and

the earliest due date. Choosing the right sequence to reach

multiple production goals is challenging in view of the conflicting objectives. It is known that choosing parts with the

shortest processing time usually gives the best throughput.

For this batch, the part with the shortest processing time (part

type B) has the lowest tardiness penalty. However, it may be

beneficial to start the process by choosing the part with the

highest penalty in order to avoid high production costs. The

fuzzy methodologies presented look at these aspects and find

a middle ground so that all the objectives can be fulfilled as

far as possible.

123

428

programmed logic

Input customer specified data: part #, number of parts, sequence in the case of sequential processing,

and due date/penalty data for parts. (1)

Input production data: idle rate of machines, and fixed robot move times within the cell. (2)

Priority-Loading (5)

Priority-unloading (6)

Priority-Loading

SPT

WEDD

Fuzzy-Machine (FM)

Fuzzy-Job (FJ)

ALL

SPT-Sequence jobs in

increasing order of their

processing times (PT)

among the many

combinations of job(j) and

machine (m). Implement

eqn.(1).

evaluating the overall contribution of the

fuzzy membership function of the part

type to the optimal performance of the

system by using equations 6 through 12

Priority-unloading

WEDD-Sequence jobs in

increasing order of the

ration of the due date of the

job(DD) to the penalty of

the job (P) assigned when

the job is late.

Implement eqn.(2).

evaluating the overall contribution of the

fuzzy membership function of the part

type to the optimal performance of the

system by using equations 13 and 14 and

12.

Percentage of improvement

of fuzzy over SPT & WEDD

WEDD, FM, and FJ individually and

pick the optimal solution and store the

data.

measure of criterion obtained from the strategies

difference of = 100 selected

improvement

measure of criterion obtained from the strategies

to which comparison is made

80%

60%

40%

20%

0%

Machine idle

cost

Tardiness cost

Throughput

time

Criteria

Improvement of fuzzy over SPT

Fig. 2 Percentage difference of improvement of fuzzy-job and fuzzymachine with respect to SPT and WEDD for batch #2

the resulting improvement when using fuzzy-job and fuzzymachine over SPT and WEDD. The bar graph compares

the performance of the different strategies. The percentage

difference of improvement is defined as given below. For

this particular case, the strategies selected are fuzzy-job and

fuzzy-machine, and the strategies to which comparison is

made are SPT and WEDD.

123

For this batch, fuzzy-job and fuzzy-machine produce identical performance, and the figure presents the results using

them as base line. In this example, SPT chooses part type B

first. Part type B has the smallest processing time, and the

highest due date over penalty ratio. At the same time, SPT

tends to leave part type A for processing to the end. Part type

A has the highest penalty; therefore, tardiness cost would tend

to be higher (Fig. 2). On the other hand, WEDD chooses part

type A, which has the lowest due date over penalty ratio.

However, that part type has a very long processing time,

which at the end gives rise to a long robot idle time, and

thus a longer throughput time. This produces a very high

tardiness cost. Fuzzy-job and fuzzy-machine perform 71%

better than WEDD in this case. Although, it is expected that

WEDD would tend to give the best results when it comes to

tardiness criterion, it did not turn out to be so since WEDD

does not take into account processing times.

429

12%

10%

8%

6%

4%

2%

0%

Machine idle

cost

Tardiness cost

Throughput

time

Criteria

Fig. 3 Percentage difference of improvement of fuzzy-job and fuzzymachine with respect to WEDD for batch #3

3500

2800

2100

1400

700

0

FUZZY-J FUZZY-M

LOAD

SPT

WEDD

1600

1200

800

400

0

FUZZY-J FUZZY-M

LOAD

UNLOAD

Throughput time (s)

unloading priority for batch #3

in Table 3. In this case, part type A has the longest due date

and the highest tardiness penalty. Part type B has the smallest processing time, while part type C has values in between,

and the shortest processing time (machine M3). Results for

this batch are shown in Fig. 3. Fuzzy-job and fuzzy-machine

find a middle ground by initially choosing part type C. SPT

chooses part type C as well, since it has the lowest processing time. Results show the same performance for fuzzy-job,

fuzzy-machine and SPT. There is a 10.7% improvement over

WEDD in terms of tardiness cost.

In order to study whether further enhancement can be

obtained by any of the different strategies, further analysis was done using an unloading priority. In some cases,

when using an unloading priority, results tend to be better.

The results cannot be generalized. For illustration purposes,

results for batch #3, using an unloading priority are shown

better results than loading priority. In order to verify whether this is true, three more tests using data shown in Table 6

labeled as batch #4, 5, and 6 were analyzed. Once again,

fuzzy-job and fuzzy-machine are compared against SPT and

WEDD, but taking into account loading and unloading priority. There are eight combinations (methodology-priority) in

total. In addition, when parts have the same processing time,

comparisons are also made to a relevant study (Balakrishnan

et al. 2001). Results are shown in Figs. 57. The tables and

figures indicate the improvement in percentage of the best

performance compared to the worst, and the improvement of

the best methodology compared to SPT.

The results pertaining to data (batch#4 in Table 6) are

shown in Fig. 5. It is clear that fuzzy-job outperforms the

other methodologies. For fuzzy-methodology, there is also

a remarkable similarity between the results of loading and

unloading priority, except in terms of tardiness cost. Although

unloading priority is 1.8% better than loading priority in tardiness cost criterion, the percentage difference is too small

to conclude that the fuzzy-job with unloading priority yields

the best results. In terms of robot idle time, there is not much

difference in performance between the priorities. For this

specific case, comparisons against Hathouts heuristics are

possible since processing times are equal. However, these

comparisons are only possible in terms of throughput since

Hathouts problem does not take into account other criteria.

The throughput time for an identical batch of parts with loading priority produces a 7.5% improvement for fuzzy-job strategy in comparison to Hathouts. No consistent pattern could

be observed in regards to the strategy that performed worst.

To tal tardiness cost ($)

Percentage of improvement

of fuzzy over WEDD

SPT

WEDD

UNLOAD

700

650

600

550

500

FUZZY-J FUZZY-M

LOAD

SPT

WEDD

UNLOAD

123

430

Percentage

improvement in

performance

Batch #4

A

B

C

Batch #5

A

B

C

Batch #6

A

B

C

# Of parts

Job #

M1

M2

M3

M4

DD(p)

P(p)

PT(p)

TP(p)

C(p)

3

2

2

1,2,3

4,5

6,7

15

0

15

18

18

16

0

0

0

0

12

0

120

140

130

4

2

3

99

60

62

30

70

43.3

3

4

4

4

5

3

1,2,3,4

5,6,7,8,9

10,11,12

85

0

0

85

85

0

0

0

85

0

85

0

840

870

760

1

2

3

680

850

255

840

435

253.3

3

4

1

4

5

3

1,2,3,4

5,6,7,8,9

10,11,12

48

0

0

43

41

0

0

0

45

0

49

0

440

470

310

1

2

3

364

450

135

440

235

103.3

3

4

1

Percentage

improvement in

performance

Part type

10%

0%

Machine idle cost

Tardiness cost

FUZZY-J UNLOAD

40%

20%

0%

Tardiness cost

FUZZY-J LOAD

SPT

HATHOUT'S

Throughput time

Criteria

Throughput time

Criteria

FUZZY-J LOAD

60%

FUZZY-J UNLOAD

SPT-LOAD

SPT-UNLOAD

Percentage of

improvement over the

worst performance

80%

60%

40%

20%

0%

Machine idle cost

Tardiness cost

Throughput time

Criteria

FUZZY-M LOAD

FUZZY-M UNLOAD

SPT

better than loading priority except in tardiness cost criteria.

An important factor to be noticed is that, unloading is better

(>6%) than loading priority in all cases. Nevertheless, when

loading is better in tardiness criteria, the difference is 23%.

Thus, SPT and fuzzy-machine with loading priority have the

best performance.

HATHOUT'S

Summary of results

the best performance as shown in Fig. 6. As in the previous

example, loading and unloading priority give similar results.

Loading performance is better by 35.3% for tardiness cost.

In terms of throughput, unloading and loading have the same

performance; however the robot idle time for the unloading

priority is longer, which means less robot-related production cost. The difference is 0.8%. This difference is minimal;

hence, it can be concluded that fuzzy-machine with loading priority has the best performance. For this specific case,

comparisons against Hathouts heuristics are also possible

since all processing times are equal to 85 seconds. In terms

of throughput, Hathouts heuristics gives identical results to

those obtained using fuzzy-machine.

For the next case, batch#6 given in Table 6, SPT and fuzzymachine have the best performance, as shown in Fig. 7.

123

Figs. 810 show the percentage of how many times the methodology-priority combination was the best in the thirty trials.

For compactness, letters represent the combination methodology-priority. For example, FJ means fuzzy-job, FM means

fuzzy-machine, W and S mean WEDD and SPT respectively.

Likewise, L and U represent loading and unloading priority,

respectively.

The first aspect to analyze is the performance of the methodologies with respect to the machine idle cost criterion

(Fig. 8). Fuzzy-job has the best performance. It is best 56.6%

of the time with unloading priority, and 53.3% of the time

with loading priority. SPT and WEDD with loading priority

have the worst performance. Unloading priority seems to

yield somewhat better results than the loading priority.

For the tardiness cost criterion (Fig. 9), fuzzy-job with

unloading priority has the best performance followed by

Percentage of times a

methodology-priority is best

431

60%

50%

40%

30%

20%

10%

0%

FJ-L

FJ-U

FM-L

FM-U

W-L

W-U

S-L

S-U

S-L

S-U

S-L

S-U

Methodology-priority

Percentage of times a

methodology-priority is

best

60%

50%

40%

30%

20%

10%

0%

FJ-L

FJ-U

FM-L

FM-U

W-L

W-U

Methodology-priority

Percentage of times a

methodology-priority is

best

60%

50%

40%

30%

20%

10%

0%

FJ-L

FJ-U

FM-L

FM-U

W-L

W-U

Methodology-priority

unloading priority would perform best in terms of the tardiness criterion since jobs would be unloaded faster, thus

reducing tardiness costs. However, this is not always the case.

For instance, the loading priority performs better for fuzzymachine and SPT. WEDD with the loading priority showing

the worst performance, followed by SPT.

In terms of throughput criterion (Fig. 10), results are very

similar to before, fuzzy-job with unloading priority showing

the best performance. SPT and WEDD with loading priority

have the worst performance. In general, fuzzy-job with the

unloading priority gives the best results; SPT and WEDD

show poor performance. Furthermore, the unloading priority usually performs better than the loading priority, except

for the tardiness cost criterion, where results are difficult to

generalize.

Conclusions and recommendations

The performance of the proposed fuzzy logic based methodologies is very promising. They have shown much better performance than traditional dispatching rules such as SPT and

and WEDD may still be good when considering single objectives such as maximizing the throughput time or minimizing

the tardiness cost, respectively. This work has shown that

fuzzy methodologies are able to combine several objectives

for effective scheduling of jobs. The results presented also

show that fuzzy-job is more effective than fuzzy-machine.

As indicated before, the difference in performance can be

attributed to the way each strategy analyzes the contribution

of the jobs to reach the objectives. Fuzzy-job considers the

attributes of the job only, while fuzzy-machine evaluates the

contribution of the job-machine combination. The enhancement in performance shown by the fuzzy-job comes from

analyzing each job, keeping in perspective all the machines

and their ability to process a set of jobs that constitutes a

batch. It does not restrict the analysis to just a job-machine

combination. The results also indicate a slight difference in

performance between SPT and WEDD. SPT has a tendency

to perform better in machine idle cost and throughput time

criteria, while WEDD performs better in tardiness cost criterion.

For the tardiness cost criterion, the unloading priority has

proved to be the best priority for fuzzy-job and WEDD,

while the loading priority has been the best choice for fuzzymachine and SPT. In regards to improving the machine idle

cost and throughput time, unloading priority turns out to be

the best priority. The methodologies were successfully implemented in an automated machine cell. The two strategies

performed quite well and the results obtained from simulation (off-line) show only a marginal difference with those

from actual implementation (online). The slight difference

is unavoidable due to the preset control architecture of the

robot and communication aspects.

The capability of the custom designed software used to

evaluate the performance of the two strategies can be effectively used for simulations (off-line) and actual implementations (online). The software has the ability of producing

simulation in text or graphic mode for valuable data collection for further studies. An example is the generation of

different sequences for a wide variety of batches for SPT,

WEDD, fuzzy-job and fuzzy-machine methodologies, or all

of them. When the sequences of all the methodologies are

displayed, comparisons of the performance of the strategies

for each objective can be easily seen.

The methodology developed has been successfully implemented on a robot based FMC in the research laboratory. The

microprocessors of industrial robots are not designed to perform any computations other than what is required for motion

control of robot. Hence an external PC was employed for performing all the computations required for the methodology

developed. The algorithm implemented on a personal computer generates the best strategy and schedule for the selected data set. The robot programs corresponding to loading

123

432

PC based input/output card which calls the appropriate programs in the correct sequence. The PC based cell controller

performed quite well and clearly demonstrates that it will be

possible to implement the developed algorithm without any

modifications to the cell architecture. The algorithm is extremely fast and has made it possible to implement the new

fuzzy-based methodology for near real-time control of an

FMC. The effectiveness of fuzzy-job in improving throughput in a multi-objective environment could be further enhanced by automatically assigning weights to the objectives. A

study of this nature would provide much better insight on the

impact of membership functions on the quality of solutions.

This is currently investigated as part of an on-going study.

The use of an adaptive algorithm is being explored that determines automatically the weights for various objectives and

their impact on the quality of solution. This assignment could

be complex since there are four membership functions, which

could be, assigned ten weight values ranging from 0 to 1. The

number of possible combinations would be enormous (104 ).

There are two possible ways of assigning and evaluating these

weights. The first one could be by using rules, and would

require substantial user input and further experiments. A second approach, which may be more efficient, is by using techniques such as genetic algorithms or neighborhood search

methods. These approaches may require substantial computation times, and the improvement in performance will have

to be evaluated against the computational time that would be

needed. Studies are also currently underway in developing

a genetic algorithm based approach and the results will be

made available at the conclusion of the study.

Although non-sequential methodology proved to be more

efficient than sequential mode, further research can be done

to enhance the performance of non-sequential mode. In the

present work, no special techniques were utilized to check

the movements of parts in non-sequential method. An objectoriented function that checks for presence of no conflicts in

part allocation was utilized. However, by using intelligent

techniques with look-ahead features, these conflict-checks

can be further enhanced and results for non-sequential mode

may show further improvement.

The proposed methodology can be applied to a wide variety of manufacturing environments. For example, automated

printed circuit board assembly is a clear example of a process

that could benefit from an optimized selection of a combination of sequential and non-sequential processing. A robot

assisted automated cell for manufacture and assembly of

plastic valves with several sub-components is currently being

developed and the industry is keen in applying the proposed

methodology for controlling the robot movements within the

cell. This process can be categorized as sequential processing. This will provide a great opportunity for implementing

the developed control strategy in a real-world application.

123

References

Agentis, A. (2000). Scheduling no-wait robotic cells with two and three

machines. European Journal of Operational Research, 123, 303

314.

Agentis, A., & Pacciarelli, D. (2000). Part sequencing in three machine

no-wait robotic cells. Operations Research Letters, 27, 185192.

Balakrishnan, S., Hathout, L., & Popplewell, N. (2001). Sequential versus Non-sequential loading and influences on Intra-Cell sequences.

In Proceedings of the IASTED International Conference, Robotics

and Applications (pp. 95100). ISBN:0-88986-313-X.

Bellman, R. E., & Zadeh, L. A. (1970). Decision-making in a fuzzy

environment. Management Science, 17, 141164.

Brauner, N., & Finke, G. (2001). Optimal moves of the material handling

system in a robotic cell. International Journal of Flexible Manufacturing Systems, 74, 269277.

Chen, S., & Lin, L. (1999). Reducing total tardiness cost in manufacturing cell scheduling by a multi-factor priority rule. International

Journal of Production Research 37(13), 29392956.

Chen, H., Chu, C., & Proth, J. (1997). Sequencing of parts in robotic

cells. The International Journal of Flexible Manufacturing Systems,

9, 81103.

Cox, E. (1998). The fuzzy systems handbook (2nd ed.). New York: AP

Professional.

El-Bouri, A., Balakrishnan, S., & Popplewell, N. (2000). Sequencing

jobs on a single machine: A neural network approach. European

Journal of Operational Research, 126, 474490.

Holthaus, O., & Ziegler, H. (1997). Improving job shop performance by

coordinating dispatching rules. International Journal of Production

Research 35(2), 539549.

Jain, A. S., & Meeran, S. (1996). Scheduling a job-shop using a

modified back-error propagation neural network. In Proceedings of

the First Turkish Symposium on Intelligent Manufacturing Systems

(pp. 3031, 462474).

Jawahar, N., Aravindam, S. G., Ponnambalam, S. G., & Karthikeyan, A.

A. (1998). Genetic algorithm-based scheduler for setup-constrained

FMC. Computers in Industry, 35, 291310.

Kazerooni, A., Chan, F. T. S., & Abhary, K. (1997). Real-time operation

selection in an FMS using simulationa fuzzy approach. Production

Planning & Control, 8, 771779.

Kutanoglu, E., & Sabuncuoglu, I. (1999). An analysis of heuristics in a

dynamic job shop with weighted tardiness objectives. International

Journal of Production Research, 37(1), 165187.

Moreno, A., & Ding, F. (1993). A constructive algorithm for concurrently selecting and sequencing jobs in an FMS environment. International Journal of Production Research, 31(5), 11571169.

Restrepo, I. M., & Balakrishnan, S. (2001). Fuzzy logic based robot controller for intelligent scheduling. In Proceedings of the 32nd. International Conference on Applied Modeling and Simulation (Vol. 2,

pp. 358363).

Sannomiya, N., & Iima, H. (1996). Application of genetic algorithm to

scheduling problems in manufacturing processes. In Proceedings of

the IEEE Conference on Evolutionary Computation (pp. 523528).

Sethi, S. P., Chandrasekaran, R., Drobouchevitch, I., & Sriskandarajah,

C. (2004). Scheduling multiple parts in a robotic cell served by a

dual-gripper robot. Operations Research, 52, 6582.

Yalcin, A., & Boucher, T. (1999). An architecture for flexible manufacturing cells with alternate machining and alternate sequencing. IEEE

transactions on Robotics and Automation, 15(6), 11261130.

Vidyarthi, N. K., & Tiwari, M. K. (2001). Machine loading problem

of FMS: a fuzzy-based heuristic approach. International Journal of

Production Research, 39(5), 953979.

Yager, R. (1978). Fuzzy decision making including unequal objectives.

Fuzzy Sets and Systems, 1, 8795.

Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8, 338353.

- bitszg553mid5Uploaded byShivam Shukla
- Operating_IIUCUploaded byZishan Cse
- Blueprint TrainingUploaded byTong Bbm
- CPU_SchedUploaded byFazrul Rosli
- Logica FuzzyUploaded byLivia
- Available to promise ATPUploaded byAlexey Lizun
- A Historical Review of Evolutionary Learning Methods for Mamdani-type Fuzzy Rule-based Systems_ Designing Interpretable Genetic Fuzzy SystemsUploaded byjkl316
- Recent Neutrosophic Models for KRP SystemsUploaded byMia Amalia
- Simulation of Scheduling AlgorithmsUploaded byashishpatel_99
- Release TimeUploaded byAbdul Wadood
- SMAW Process Parameters Optimization using Taguchi & Fuzzy LogicUploaded byAnonymous kw8Yrp0R5r
- A framework for PEST analysis based on neutrosophic cognitive map: case study in a vertical farming initiativeUploaded byMia Amalia
- Operating SystemUploaded byHeran Abebe
- QTUploaded byManan Mehra
- 00816520Uploaded byRadhika Vadlapatla
- IJGCA 030405Uploaded byijgca
- Thread Scheduling and DispatchingUploaded bypankaj_sharma2222704
- pizzzUploaded byQuangdat To
- Introduction to Fuzzy SetsUploaded byKathiresan Manickam
- cpu_sched.Uploaded bySushobhan Bhaumik
- Sporadic SystemUploaded byvivek2585
- Assignment 2 w SolUploaded byMarcel Caday
- Iar Rtos ArticleUploaded bykuku2008
- Dbms Important Questions Unit Wise With AnswersUploaded bysameeksha chiguru
- 3348-8095-1-PBUploaded byerum
- A Deadline and Cost Constrained Optimization Algorithm for Scheduling Applications in Grids Based on Proportional Share SystemsUploaded byMing Geng
- RTS1Uploaded byASHA KAINDAL
- 6.0 Fuzzy SetsUploaded bymir1973
- Real-Time Scheduling Strategy for Wireless Sensor Networks O.SUploaded byijdps
- 17512-2018-Winter-model-answer-paper.pdfUploaded byJayesh Phale

- prilog6preporucenisadrajstudijeUploaded byInesGalic
- Zmegac Strast 005Uploaded byInesGalic
- 276 E-marketing - SadrzajUploaded byInesGalic
- Evolucija SindikataUploaded byInesGalic
- Stvarno PravoUploaded byInesGalic
- Linux BiblijaaUploaded byzahariev
- Knjiga Udjela d.o.o.-primjerUploaded byInesGalic
- Vlak u Snijegu Mato LovrakUploaded bydamec1234
- Propisi TkUploaded byMajda Kadrić Kamberović
- 169223825 Organizacijska KulturaUploaded byInesGalic
- Pouzdanost komponentiUploaded byAleksa Popovic
- 1_195Uploaded byInesGalic
- Pravilnik o Sadrzaju i Nacinu Vodjenja Trgovacke KnjigeUploaded byDuško Laštro
- Prav o Obl Sad i Nac Vodj Trgov Knjige 91 15Uploaded byInesGalic
- Kako Umetnuti Sadraj WEBUploaded byNeko Plus
- ccna_sem1_v30_mod1Uploaded byzgenzlon
- fulltext 11Uploaded byInesGalic
- Platonova Teorija OdgojaUploaded byInesGalic
- Internet i WWWUploaded byAladin Hadžić
- EKSPERTNI SUSTAVIUploaded byInesGalic
- Power Point 2000Uploaded bybpajic
- 11 Juric Racic RadicaUploaded byInesGalic
- CIMBENICI_PROFESIONALNOG_STRESAUploaded byMerisa Mahmutovic
- Frontpage CourseUploaded byInesGalic
- 4_ANTONICUploaded byInesGalic
- Kratko Uputstvo Za Vezivanje KravataUploaded bybluemalena
- The Complete Poems of Edgar Allen PoeUploaded byInesGalic
- agileProcessUploaded byfentacool
- agileProcessUploaded byfentacool
- agileProcessUploaded byfentacool

- TCS Verbal Set 2Uploaded byAtanu Brainless Roy
- Multipathing in Aix With Netapps StorageUploaded byAnbao Cheng
- COMPUTER VIRUS.pptxUploaded byJsmBhanot
- ANU Key Dates WestpacUploaded bymrfutsch
- Disaster Management (NDMA)Uploaded byMuhammad Imran
- http7Uploaded byRafiaZaman
- IT0322 Web Systems LAB ManualUploaded byvijay81541
- Opportunity Assessment and the Entrepreneurial ProcessUploaded byMD Salahuddin Kader Monna
- 4.1 BoilerUploaded byrashm006ranjan
- May 2011 - Pmwt Featured Paper - Al-maghraby - Pmo - Final2Uploaded byhazmy
- Aerodynamics Lab 1Uploaded byBassam Riachi
- Connecting the Breakout Board to the Sun Harvester Shield V1.0Uploaded byBunPământ
- Electronics for You 201310Uploaded byMargarita Alcevska
- Final Summer Training Report Mohit PalUploaded bybharat sachdeva
- 10 ErrorUploaded byprogressksb
- Exercise compUploaded byatulsemilo
- 3a action plan yr 9 websiteUploaded byapi-348858341
- GSM HLR ArchitectureUploaded bySreeram Subramaniapillai
- Camworks Mill & Turn Tutorial 2003Uploaded bytabibkarim
- ch07Uploaded byabid
- Akhlaq AhmedUploaded byDeny P. Sambodo
- ME2303– DESIGN OF MACHINE ELEMENTSUploaded byGowtham Raj
- Volkswagen Group CSR IATF 16949 Sept 2017Uploaded bySelvaraj S
- Schematic 420EUploaded byEVCY
- Characterization of Zeolite-Based Coatings For Adsorption Heat PumpsUploaded byFreone
- Solution 3Uploaded byHerman Komba
- Ict94rjb - Thermoelectric Heat SinkUploaded byCipta Andri Jhona Sinuraya
- Failed Fuel Experience Mechanism and ManagementUploaded byRajib
- sao evidence programmingUploaded byapi-376826833
- Consider the Following SentencesUploaded byMehar Chand