Information Technology Reference
In-Depth Information
generation quantity
M
is a fixed positive integer, and maximum event generation
probability
P
is a fixed real number with a [0
,
1] interval. These two parameters
represent how often problems occur in the plant. They are respectively decided
based on the plant size and the plant equipment degradation rate. In this model,
the plant environment generates an event with a probability
r
P
.Thisevent
generation process is iterated
M
times every time step. In other words, the
higher the production rate, plant size and plant equipment degradation rate,
the more problems that occur. This represents our assumption that the amount
of problems with risks is proportional to three factors, the production rate, the
plant size, and the plant equipment degradation rate. The generated event is
assigned to one randomly selected agent. This represents when the plant worker
finds equipment degradation that should be repaired.
×
3.3 Event
An event represents an incident that can grow from a risk to a fatal accident.
One event
e
has the following parameters:
-
time
t
e
,
-
reduction rate
R
e
,
-
required skill
S
e
,and
-
required workload
w
e
.
Time
t
e
is the number of steps passed after generation of this event without
processing by the worker agents. Reduction rate
R
e
is a fixed real number with
a[0
,
1] interval. Suppose that
E
is a set of events that are being processed or
not processed yet, the production rate
r
is calculated by using the following
equation:
r
=
e∈E
(1
−
R
e
)
.
(1)
This means, the more problems that remain, the less the production rate. Re-
quired skill
S
e
is
L
-dimensional vector
S
e
=
.Thenumber
of dimensions
L
represents the number of skill types required in the plants. Each
element of the vector represents the skill necessary to handle the equipment and
solve problems. Required workload
w
e
, which is a positive integer, is reduced
by 1 when the worker agent processes event
e
.When
w
e
becomes 0, event
e
is
removed from the simulations. This means the plant worker solved the problem.
{
s
i
|
1
≤
i
≤
L,s
i
∈{
0
,
1
}}
3.4 Worker Agent
The worker agents represent the plant workers. They process the events using
their skills. The skills of a worker agent is represented by
L
-dimensional vector
S
w
=
, which is the same as required skill
S
e
.
In this implementation, the worker agents form a hierarchical structure. They
can be classified as a ”manager” or ”operator.” The managers have relationships
{
s
j
|
1
≤
j
≤
L,s
j
∈{
0
,
1
}}
