Majalah Ilmiah UNIKOM

Vol.10 No. 1

71

H a l a m a n

area value

is defined for the floors

where a call (hall and car) has happened.

Figure 5. Example of the area value

We can see that

is a fixed value

and

are estimated values.

However, the value

is calculated

whenever a call is generated. Therefore the

determination of the value

is important in

the hall call assignment method.

If the value

is big, the possibility that

the elevator close to the relevant floor can

be selected is increased. Consequently, the

transportation load may be assigned to a

specific elevator which is in the area.

Furthermore, the average waiting time

increases and the total running frequency of

elevator decreases. Therefore the power

consumption is reduced. If

is small, the

average waiting time decreases and the

running frequency increases.

The predefined area-weight is usually

used and it is defined according to the

traffic modes classified by the passenger’s

traffic pattern. In this study, we use the fuzzy

approach to model the uncertain situation of

the system and to determine the area-

weight.

IV. A TYPE-2 FUZZY MODEL TO

DETERMINE THE AREA-WEIGHT

The type-2 fuzzy model is used to

determine the area-weight. Fuzzy knowledge

is generally formulated in the form of rules.

Fuzzy Rules

We classify the facts related with the

determination of area-weight into two

groups. The first group includes the up-going

and down-going traffic amount named as

pre-area fuzzy. The second group includes

the average waiting time, long wait

probability and power consumption named

as area fuzzy.

The passenger’s traffic varies from hour to

hour and we can see that it is classified into

some basic patterns.

up-peak pattern

This pattern is the case of which the up-

going traffic is very large but the down-

going traffic is very small.

down-peak pattern

In this pattern, the down-going traffic is

much more than the up-going traffic

normal pattern

The case of which both up-going and

down-going traffic are approximately

equal.

Figure 6. Membership Function of Up-Peak Traffic