Page - 230 - in Adaptive and Intelligent Temperature Control of Microwave Heating Systems with Multiple Sources
Image of the Page - 230 -
Text of the Page - 230 -
A. Appendix
fo(z o
j(k)) =z o
j(k), thecompletedupdateequation is
woj,i(k+1) =w o
j,i(k)−η ·δoj(k)xLi (k),
=woj,i(k)−η ·δoj(k)xLi (k). (B.23)
Case2: Forweightsonlyconnectingtohiddenlayers
N1,1
N1,2 N2,1
N2,2
N2,3 N3,1
Input Layer Hidden Layer
Output Layer
YNN,1
u1(k)
u2(k)
W13,4
Bias = 1 N3,2 YNN,2
W22,4
Bias = 1 W11,2
FigureB.2. Errorflowsfromhiddenlayers to theoutput layer.
When both ends of the link are connected to nodes in hidden layers,
the update of its weight becomes complicated. There is no specified
desiredoutputfornodesinhiddenlayerstodirectlycalculatetheerror(
YNN,j(k)−Yd,j(k)
)
. Inaddition,allnodes in theoutput layerhaveto
be considered because the error flows from the node in the hidden
layer to all nodes in the output layer, such as indicated by the blue
arrows infigure B.2.
Using a similar equation as in the first case to calculate ∆woj,i(k), for
theL-th layer there is
∆wLj,i(k) =−η ∂J(k)
∂wLj,i(k) = −η ∂J(k)
∂zLj (k) ∂zLj (k)
∂wLj,i(k)
=−η ·δLj (k)xL−1i (k). (B.24)
230
back to the
book Adaptive and Intelligent Temperature Control of Microwave Heating Systems with Multiple Sources"
Adaptive and Intelligent Temperature Control of Microwave Heating Systems with Multiple Sources
- Title
- Adaptive and Intelligent Temperature Control of Microwave Heating Systems with Multiple Sources
- Author
- Yiming Sun
- Publisher
- KIT Scientific Publishing
- Location
- Karlsruhe
- Date
- 2016
- Language
- English
- License
- CC BY-SA 3.0
- ISBN
- 978-3-7315-0467-2
- Size
- 14.8 x 21.0 cm
- Pages
- 260
- Keywords
- Mikrowellenerwärmung, Mehrgrößenregelung, Modellprädiktive Regelung, Künstliches neuronales Netz, Bestärkendes Lernenmicrowave heating, multiple-input multiple-output (MIMO), model predictive control (MPC), neural network, reinforcement learning
- Category
- Technik