< prev

Page 1Page 2Page 3Page 4Page 5Page 6Page 7Page 8Page 9Page 10Page 11Page 12

Page 9 of 12
next >

Majalah Ilmiah UNIKOM

Vol.10 No. 1

73

H a l a m a n

Figure 11. Membership Function of Long Wait

Probability

k

Table 2. Area Fuzzy Rules

Fuzzy Model

We propose a fuzzy model to determine

the area-weight through a two step fuzzy

inference. Two step inference mechanism

improves the system’s stability from external

accidents and reduces the complexity of the

system. Figure 13 shows the two step fuzzy

inference mechanism.

Figure 13. The Fuzzy model to determine the area-

weight

In step 1 of the fuzzy inference engine, the

predetermined

area-weight

(

)

is

calculated using the up-going (UP) and down

-going (DN) traffic. In step 2, the adjustment

value

is determined through the fuzzy

inference mechanism using the average

waiting time (AWT), the long wait probability

(LWP) and the power consumption values

(PC). This value is added to the

predetermined area-weight.

V. SIMULATION RESULTS AND DISCUSSION

Performance of Type 2 Fuzzy Controller is

investigated using simulation studies. The

developed controller was compared with

Type-1 Fuzzy Controller. The condition of

simulation are shown in the Table 3.

According to the traffic pattern, the

simulation situation is divided into several

periods such as before lunch time (12:00 –

12:40), after lunch time (12:40 – 13:20)

and common time (13:20 – 15:00). The

evaluation criteria is the means of the

average waiting time (AWT), power

consumptions (PC) and average long wait

probabilities (LWP). In the simulation, the

Muhammad Aria

Inference Rules

Weight

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1

k

1