Information Technology Reference
In-Depth Information
6. Make public filtering to the new states, then send the residual states to the ACTIVE
table.
Following states are excluded by public filtering:
• No change state: such as I
1
, I
4
, I
7
• Cycle state: new state is same to a former state
• Divergent state: one parameter value is
¯
, which means the current time point
is the finish point.
Here we take the qualitative simulation of process of throwing ball up as
example to explain QSIM algorithm, Assuming that the height of the ball is
Y
,
speed is
.
Known constraint relationship are
DERIV(
V
and acceleration is
A
Y,V
)
DERIV(
V,A
)
)
g
A
(
t
<0
Ball moves up in initial state(t
0
,t
1
):
QS(
A,t
0
,t
1
)<
g
, std>
QS(
V,t
0
,t
1
)<(0,
¯
),dec>
Y,t
0
,t
1
)<(0,
¯
),inc>
Make all possible transformation to each parameter. I transformation is
needed because it is in the time interval now:
QS(
A
I
1
<g, std>
¼
<g, std>
V
I
5
<(0,
¯
),dec>
¼
<0,std>
I
6
<(0,
¯
),dec>
¼
<0,dec>
I
7
<(0,
¯
),dec>
¼
<(0,
¯
),dec>
<l
*
,std>
I
9
<(0,
¯
),dec>
¼
Y
I
4
<(0,
¯
),inc>
¼
<(0,
¯
),inc>
<l
*
,std>
I
8
<(0,
¯
),inc>
¼
Following creates tuple set according to constraint. Make consistent filtering
to single constraint at first. Those filtered here are denoted by C. Then combine
