Artificial Neural Network
Artificial Neural Network
December 2, 2010
CONTENT
• Introduction
• Theory
• Application
• Task and Discussion
INTRODUCTION
• Can you develop mathematical
model to integrate all parameters?
INTRODUCTION
• Or this
konveksi radiasi
matahari
konveksi Greenhouse
cover
udara
dalam ventilasi
radiasi gel. panjang
Permukaan lantai
tanah konduksi
INTRODUCTION
• Alternative solution: Artificial
Neural Network (ANN) model
• ANN : a mathematical model or
computational model that is inspired
by the structure and/or functional
aspects of biological neural
networks.
• Introduced by Waren McCulloch and
Walter Pits in 1943
INTRODUCTION
• Biological neural and artificial neural
ANN BNN
Node Cell body
Input Dendrites
Output Axon
Weight Synapses
INTRODUCTION
• Advantages of ANN:
– A neural network can perform tasks that a
linear program can not.
– When an element of the neural network fails,
it can continue without any problem by their
parallel nature.
– A neural network learns and does not need to
be reprogrammed.
– It can be implemented in any application.
– It can be implemented without any problem.
INTRODUCTION
• Disadvantages of ANN:
– The neural network needs training to
operate.
– The architecture of a neural network is
different from the architecture of
microprocessors therefore needs to be
emulated.
– Requires high processing time for large
neural networks.
INTRODUCTION
• How does ANN work in control
tech?
d4
d1
r + e u + + y
C(s) P(s)
-
controller d2 process
d3
• Algoritma Perceptron NN
– Kasus X1 or X2 X0 X1 X2 Yt
1 0 0 0
1 0 1 1
1 1 0 1
1 1 1 1
1 0 0 0
1 0 1 1
1 1 0 1
1 1 1 1
1 0 0 0
Theory
• Hasil Perceptron NN
Theory
• Feedback/recurrent
– Backpropagation NN
– Competitive networks
– Self-organizing maps
– Hopfield networks
– Adaptive-resonanse theory models
Theory
• Backpropagation NN
h0=1
xI hJ
Hidden layer
h _ in j = ∑ X iWi j
i
Theory
• Algoritma Backpropagation NN
– Langkah 2 :Feedforward
• nilai Hi dihitung dengan menggunakan fungsi
sigmoid sebagai fungsi aktivasi
1
Hi = f (h _ in j ) = − h_ in j
1+ e
Theory
• Algoritma Backpropagation NN
– Langkah 2 :Feedforward
• Unit output (Y) menjumlahkan sinyal –
sinyal input terbobot :
y _ in = ∑ H jV jk
j
∆V j k = ηδ k H i
Theory
• Algoritma Backpropagation NN
– Langkah 3 : Backpropagation
• Dimana η adalah konstanta pembelajaran
dan δk dihitung dengan persamaan
δ k = (Ot − Y )Y (1 − Y )
• Perbaikan pembobot (Wij) antara hidden
layer dengan input layer dihitung dengan
persamaan:
∆W ij = ηδ j Xi
• δj dihitung dengan persamaan:
δ j = H j (1 − H j )∑V jδ k
Theory
• Algoritma Backpropagation NN
– Langkah 3 : Backpropagation
• Pembobot baru dihitung dengan persamaan
w j (baru ) = w j (lama ) + ∆w j