modelling, analyzing and control of des and deds
TRANSCRIPT
![Page 1: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/1.jpg)
1
Modelling, analyzing andcontrol of DES and DEDS (Compendium of the recent research)
František Čapkovič
Presentation in University of Aizu
October 11, 2006
![Page 2: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/2.jpg)
2
Author's address
František ČapkovičInstitute of Informatics, Slovak Academy of Sciences
Dúbravská cesta 9, 845 07 Bratislava, Slovak Republic
e-mail: [email protected]
http://www.ui.sav.sk/home/capkovic/capkhome.htm
![Page 3: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/3.jpg)
3
Contents
IntroductionBasic definition of DEDSPetri nets in DEDS modellingDirected graphs in DEDS modelling
DEDS control synthesisDefinition of the control synthesis
Basic principle of the proposed control synthesismethod
![Page 4: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/4.jpg)
4
ExamplesClient-server cooperationTwo agents cooperation
AdaptivitySimple example
Petri nets in problem solvingProblem solving and causalityPetri nets & reachability graphs in problem solving
![Page 5: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/5.jpg)
5
Problem solving & DESSolving the DES control synthesis problemAgent-based approach to DES control synthesisExample – the maze problem solving
Approaches to solving the problemMutual intersection of autonomous solutionsSolving the global problem in the wholeUtilizing P-invariants of Petri net based model
![Page 6: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/6.jpg)
6
Modular approach to agents modellingEngineering applications
Assembly & disassembly process
Flexible manufacturing system
Conclusions
![Page 7: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/7.jpg)
7
The graphical expression of a DEDS dynamics development
![Page 8: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/8.jpg)
Causality
![Page 9: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/9.jpg)
9
Why Petri nets in DEDS modelling ????PN are able to express parallelism and conflict situationsPN can be expressed in analytical terms (in the form of the linear discrete system) as well as in the graphical formPN properties can be tested by means of the reachability tree and invariantsPN allow to use analytical approach to the DEDS control synthesisPN make possible to quantify (model) problems that
are given e.g. only verbally
![Page 10: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/10.jpg)
10
An example of a Petri net
![Page 11: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/11.jpg)
11
Parallelism
![Page 12: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/12.jpg)
12
Conflict situation
![Page 13: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/13.jpg)
13
Formal expression of the Petri net structure
Set of the PN places
Set of PN transitions
Set of PN arcs from places to transitions
Set of PN arcs from transitions to places
![Page 14: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/14.jpg)
14
Formal expression of the Petri net dynamics
Set of state vectors
Set of control vectors
Transition function
![Page 15: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/15.jpg)
15
Mathematical model of the Petri net
![Page 16: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/16.jpg)
16
A more general PN-based mathematical model of DEDS (with the multiplicity of arcs)
![Page 17: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/17.jpg)
17
Reachability tree of the above introduced Petri net
![Page 18: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/18.jpg)
18
Directed graphs (DG) in DEDS modelling
![Page 19: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/19.jpg)
19
Graph-based modelg
Tag
![Page 20: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/20.jpg)
20
State machines
Petri nets where each transition has only one input and only one output position are named state machines.They can be modelled by directed graphs (DG) without any problem.
Petri nets with general structure
In case of the general structure, when any transition is allowed to have more input positions and more output ones, the PN reachability graph has to be used.
![Page 21: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/21.jpg)
21
Petri netmodel
Reachabilitytree
Reachabilitygraph
Real DEDS
Graph model
Controlsynthesis
Transforming the PN model to the DG model
![Page 22: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/22.jpg)
22
DEDS control synthesis
Definition of the control synthesis
Control synthesis = finding the most suitable sequence of discrete events (control interferences) which is able to ensure the transition (transformation) of the system from a given initial state into a prescribed terminal state at simultaneous fulfilling control task specifications that are imposed on the control task.
Control task specifications = criteria, constraints, etc.Usually, they are not given in analytical terms. Even, often they are given only verbally.
![Page 23: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/23.jpg)
23
Basic principle of the proposed control synthesis method
Straight-lined reachability tree and the backtracking one
![Page 24: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/24.jpg)
24
Intersection of the trees yields the state trajectory(-ies)
![Page 25: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/25.jpg)
25
Procedure in analytical terms
• The staight-lined reachability tree (SLRT)
![Page 26: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/26.jpg)
26
• The backtracking (backward) reachability tree (BTRT)
![Page 27: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/27.jpg)
27
The intersection of the SLRT and BTRT
![Page 28: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/28.jpg)
28
Using the principle of causalityDue to the principle of causality any shorter feasiblesolution is involved in the longer feasible one. Hence, when
![Page 29: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/29.jpg)
29
Shorter trajectories obtained by shifting
![Page 30: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/30.jpg)
30
Example 1 – Client-server connectionThe places of the PN-based model
p1 - the client (C) requests for the connectionp2 - the server (S) is listeningp3 - the connection of C with Sp4 - the data sent by C to Sp5 - the disconnection of C by the C himself
The transitions of the PN-based model
t1, t2, t3 – discrete events realizing the system dynamics
![Page 31: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/31.jpg)
31
PN-based model and corresponding reachability tree
![Page 32: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/32.jpg)
32
Parameters of the PN-based model
![Page 33: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/33.jpg)
33
MATLAB procedure for enumerating the RT
![Page 34: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/34.jpg)
34
![Page 35: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/35.jpg)
35
Enumerated RT
Quasi-functional adjacency matrix of RT
![Page 36: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/36.jpg)
36
Space of reachable states
![Page 37: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/37.jpg)
37
Example 2 – Two agents cooperation
The agent A needs to do the activity (i.e. to solve a problem) P. However, A is not able to do P. Consequently, A requests the agent B to do P for him.
The places of the PN-based model:
p1 – A wants to do Pp2 - A waits for an answer from Bp3 - A waits for a help from Bp4 - the failure of the cooperationp5 - the satisfying cooperationp6 - A requests B to do Pp7 - B refuses to do P
![Page 38: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/38.jpg)
38
p8 - B accepts the request of A to do Pp9 - B is not able to do P
p10 - doing P by Bp11 - B receives the request of Ap12 - B is willing to do P for Ap13 - the end of the work of B
The transitions of the PN-based model:
t1 – t9 represent discrete events realizing the system dynamics
![Page 39: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/39.jpg)
39
PN-based model
![Page 40: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/40.jpg)
40
Enumerated RT
![Page 41: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/41.jpg)
41
The reachability graph
![Page 42: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/42.jpg)
42
The space of reachable states
![Page 43: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/43.jpg)
43
Control synthesis
The initial state
The terminal state – the successful cooperation
![Page 44: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/44.jpg)
44
The intersection of the SLRT and BTRT
![Page 45: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/45.jpg)
45
The state trajectories – the successful cooperation
![Page 46: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/46.jpg)
46
Graphic tool - GraSim
![Page 47: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/47.jpg)
47
Succesfull cooperation 1
![Page 48: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/48.jpg)
48
Succesfull cooperation 2
![Page 49: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/49.jpg)
49
The failured cooperation – when B is not able to do P
![Page 50: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/50.jpg)
50
Adaptivity (flexibility)
Choosing the most suitable trajectory from the feasible ones in order to adapt the system behaviourto external demands (conditions)Changing the structure of the system model in order to express more kinds of the system behaviour.
Choosing the most suitable behaviour from the feasible ones. It is illustrated in the next example.
There are two kinds of the adaptivity in the DEDS control synthesis
![Page 51: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/51.jpg)
51
Example 3 – Two processes
![Page 52: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/52.jpg)
52
{p1, p2, p3} – 1st process P1{p5, p6, p7} – 2nd process P2
p4 – the structural element that is able to influence• the mutual exclusion of P1 and P2• the sequencing of P1 and P2• the re-running of P1 and P2
![Page 53: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/53.jpg)
53
![Page 54: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/54.jpg)
54
![Page 55: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/55.jpg)
55
Exclusion of the process P2
![Page 56: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/56.jpg)
56
Three possibilities of the P2 exclusion
![Page 57: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/57.jpg)
57
Causality and problem solving
STRIPS (STanford Research Institute Problem Solver) [Fikes and Nilsson]is associated with the so called block world paradigm.
Petri nets in problem solving
b
c
a
b
c
a
initial state x0 terminal state xN
solving
table
![Page 58: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/58.jpg)
58
Expressing causality by Petri nets and reachability graphs
A fragment of PN
Full reachability graph
![Page 59: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/59.jpg)
59
Interpretation of the Petri net places wrt. the problem
p1 – a, b, c are separated on the table (T); {a,b,c}p2 – a, b are separated on T; c is on b; {a,c/b}p3 – a, c are separated on T; b is on c; {a,b/c}p4 – a, b are separated on T; c is on a; {c/a,b}p5 – b, c are separated on T; a is on c; {b,a/c}p6 – b, c are separated on T; a is on b; {a/b,c}p7 – a, c are separated on T; b is on a; {b/a,c}p8 – b is on T ; c is on b ; a is on c; {a/c/b}p9 – c is on T ; b is on c ; a is on b; {a/b/c}p10 - c is on T ; a is on c ; b is on a; {b/a/c}p11 - a is on T ; c is on a ; b is on c; {b/c/a}p12 - b is on T ; a is on b ; c is on a; {c/a/b}p13 - a is on T ; b is on a ; c is on b; {c/b/a}
![Page 60: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/60.jpg)
60
States as the nodes of the reachability treeSolution of the problem
x0 = (0,0,0,1,0,0,0,0,0,0,0,0,0)’ - the initial statex9 = (0,0,0,0,0,0,0,0,1,0,0,0,0)’ - the terminal state
The solution is
x2 x4x0 x9
where
x2 = (1,0,0,0,0,0,0,0,0,0,0,0,0)’
x4 = (0,0,1,0,0,0,0,0,0,0,0,0,0)’
![Page 61: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/61.jpg)
61
Solving the DES control synthesis problems
• Agent-based approach
• Engineering applications of AI methods• Block world paradigm
• Hanoi tower paradigm
• Control of the flexible manufacturing system
![Page 62: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/62.jpg)
62
Agent-based approach to DES control synthesis
Agent A1 Agent A2
Control Agent Ac
Knowledge
![Page 63: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/63.jpg)
63
Example 4 – Agent-based control synthesis
Problem to be solved – the maze problem
A1A2
![Page 64: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/64.jpg)
64
Problem formulationMaze consists of 5 rooms denoted as 1, 2, 3, 4, 5
connected by doors c1, c2, …, c7, c8 for A1 anddoors m1, m2, …, m6 for A2.
Initially, A1 is in the room 3, A2 is in the room 5.Doors can be open (closed) by the control agent Ac.Only door c7 is permanently open (uncontrollable).
Agent Ac observes only discrete events fromsensors built-in the doors.
The control synthesis problem is the following:Only door
![Page 65: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/65.jpg)
65
Control synthesis problem to be solved
To find the feedback controller (a policy of theagent Ac) that will fulfil the conditions:
1. A1, A2 never occupy the same room simul-taneously
2. It is always possible for both of them to return to their initial position
3. The agent Ac should enable both of them to behave as freely as possible (with respect to (1.), (2.) )
![Page 66: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/66.jpg)
66
Petri net-based model of the agents A1, A2
Agent A1 Agent A2
![Page 67: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/67.jpg)
67
Parameters of the mathematical models
Agent A1Parameters of the PN-based model
![Page 68: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/68.jpg)
68
Adjacency matrix of RG and the space of reachablestates
![Page 69: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/69.jpg)
69
Agent A2
Parameters of the PN-based model
![Page 70: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/70.jpg)
70
Adjacency matrix of RG and the space of reachablestates
![Page 71: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/71.jpg)
71
Approaches to solving the problem
I. Mutual intersection of autonomous solutions
II. Solving the global problem in the whole
III. Utilizing the invariants of the Petri nets model
![Page 72: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/72.jpg)
72
Illustrative exampleI. Mutual intersection of autonomous solutions
Agent A1Autonomous
![Page 73: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/73.jpg)
73
Agent A2Autonomous
![Page 74: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/74.jpg)
74
The final solution
Agent A1 Agent A2
![Page 75: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/75.jpg)
75
![Page 76: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/76.jpg)
76
![Page 77: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/77.jpg)
77
II. Solving the global problem in the whole
The PN-based mathematical model
![Page 78: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/78.jpg)
78
The global solution
Agent A1 Agent A2
![Page 79: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/79.jpg)
79
![Page 80: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/80.jpg)
80
![Page 81: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/81.jpg)
81
![Page 82: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/82.jpg)
82
![Page 83: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/83.jpg)
83
III. Utilizing the invariants of the Petri nets model
Additional PN places – the so called slacks – areintroduced
i.e. in our example
![Page 84: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/84.jpg)
84
Principle of the invariants method
![Page 85: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/85.jpg)
85
![Page 86: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/86.jpg)
86
slacks slacks
A1
A2
![Page 87: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/87.jpg)
87
![Page 88: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/88.jpg)
88
The final solutionIt is the same as in case II, i.e.:
![Page 89: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/89.jpg)
89
![Page 90: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/90.jpg)
90
Comparison of the approaches(I) The approach based on the intersection of the autono-
mous solutions is simple but multiplicative – i.e. the autono-mous solution has to be found for any agent. In addition tothis the process of the intersection has to be performed(II) The global approach solves the problem simultaneously,together with the interconnections among agents. It savesthe computational time, but the problem dimensionality isgreater(III) The approach based on invariants decreases (in com-parison with (II)) the RG, in spite of the fact that it increa-ses the number of the PN places (it adds the slacks)
![Page 91: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/91.jpg)
91
Modular approach to agents modelling
![Page 92: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/92.jpg)
Interpretation of PN places
p1 – A is free p11 – A is not able to help
p2 – A has to solve a problem P p12 – A is able to help
p3 – A is able to solve P
p4 – A is not able to solve P
p5 – P is solved
p6 – A contacts another agent
p7 – A asks another agent for help
p8 – A is asked by another agent for help to solve its problem
p9 – A refuses the help
p10 – A accepts the request for help
![Page 93: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/93.jpg)
93
Interface
Agent 1Agent 2
![Page 94: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/94.jpg)
94
Incidence matrix F of the PN model of MAS
![Page 95: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/95.jpg)
95
Incidence matrix G of the PN model of MAS
![Page 96: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/96.jpg)
96
The feedback control A1 is the controlled system , A2 is the controller
![Page 97: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/97.jpg)
The feedback
![Page 98: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/98.jpg)
98
Engineering applications of AI
Assembly & disassembly problem solving:
Exchange of the bad bearing B
![Page 99: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/99.jpg)
99
Utilizing the block world paradigm
![Page 100: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/100.jpg)
100
Petri net-based model
![Page 101: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/101.jpg)
101
p1 – the initial configuration p10 – C is put on E
p2 – D is disassembled p11 – D is put on C
p3 – D is put aside
p4 – C is disassembled
p5 – C is put aside
p6 – A is free of parts
p7 – B is disassembled and put aside
p8 – E is prepared for using
p9 – E is put on A
The interpretation of PN places
![Page 102: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/102.jpg)
102
The reachability graph
![Page 103: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/103.jpg)
103
The state space
![Page 104: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/104.jpg)
104
Utilizing the Hanoi tower paradigm
The initial state of the Hanoi tower puzzle
C
B
A
![Page 105: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/105.jpg)
105
The terminal state of the Hanoi tower puzzle
![Page 106: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/106.jpg)
106
Usage of this paradigmThis paradigm can be utilized in assembly when
• the space for putting parts is limited
• the assembled parts have very different mass
• the parts are fragile, etc.
![Page 107: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/107.jpg)
107
Comparing the puzzle and the assembly process
![Page 108: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/108.jpg)
108
The PN-based model
![Page 109: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/109.jpg)
109
p1 – the initial configuration p10 – B is ejected
p2 – D is put on (symbolic) peg 1 p11 – E is available
p3 – the configuration without D p12 – A is free of parts
p4 – the situation on the peg 2
p5 – the configuration without C
p6 – E is put on A
p7 – D is added
p8 – C is added
p9 – the final configuration
The interpretation of the PN-places
![Page 110: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/110.jpg)
110
![Page 111: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/111.jpg)
111
The state space
![Page 112: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/112.jpg)
The flexible manufacturing system
Consider the robotic cell with two conveyors C1, C2, the NC-machine M, with the buffer B (having the input part B1and the output part B2), and the robot R.
Defining the PN places and transitions:
p1 = waiting the input parts t1 = taking from C1 by Rp2 = waiting the output parts t2 = machining by M p3 = R is available t3 = putting on C2 by Rp4 = M is availablep5 = contents of B
![Page 113: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/113.jpg)
The PN-based model of the FMS
![Page 114: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/114.jpg)
The reachability tree and reachability graph
![Page 115: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/115.jpg)
The model parameters
![Page 116: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/116.jpg)
The RT and state space
![Page 117: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/117.jpg)
The ambiguity and how to deal with it
When the capacities of places are infinite the ambiguity occurs in the matrix A.
Namely, the cycles engender in the RT and RG. The state space of the reachable states is infinite. Infinity is expressed by the symbol
Hence, in order to find a reasonable solution, the finite capacities of the PN places have to be determined.
![Page 118: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/118.jpg)
Dealing with the ambiguity
PN model with theinfinite state space
Degenerated RT and RG
Limiting capacitiesof PN placesControl synthesis
problem to be solved
Standard PN model,RT and RG
![Page 119: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/119.jpg)
The finite capacity
![Page 120: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/120.jpg)
The finite capacities
![Page 121: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/121.jpg)
The reachability graph
![Page 122: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/122.jpg)
The state space of reachable states
![Page 123: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/123.jpg)
The control synthesis
![Page 124: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/124.jpg)
Using the system GraSim
![Page 125: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/125.jpg)
The trajectory No. 3
![Page 126: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/126.jpg)
The other trajectories
![Page 127: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/127.jpg)
Intelligent control synthesis
DEDS control task specifications are usually given in nonanalytical terms, often only verbally.
To choose the most suitable trajectory knowled-ge-based approaches must be used.
The knowledge base (KB) expressing the control task specifications in the form of IF-THEN rules can be modelled by means of the logical and/or fuzzy PN. Thus, the KB can be expressed in analytical terms.
![Page 128: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/128.jpg)
Control taskspecifications
IF-THEN rules
DEDSFeasible
trajectories
Knowledge-based choice of the trajectory
![Page 129: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/129.jpg)
129
ConclusionsSimple general method of DEDS modelling,analyzing and control synthesis was presen-tedIts applicability to different kinds of systemswas demonstratedA simple general method for agent-basedproblem solving was presentedIts applicability to solving the problem of theDES control synthesis was demonstrated
![Page 130: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/130.jpg)
130
Three different approaches to the control synthesis of the same DES were illustrated on the example:
Mutual intersection of autonomous solutions of the elementary agentsSolving the global problem in the wholeUtilizing the invariants of Petri net-based model
The approaches were compared and evaluated.
![Page 131: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/131.jpg)
and finally,
Several engineering applications werepresented to illustrate the applicabilityof the approach
![Page 132: Modelling, analyzing and control of DES and DEDS](https://reader033.vdocument.in/reader033/viewer/2022043020/626c10dc8e69190e254fe934/html5/thumbnails/132.jpg)
132
Future work on this way
To innovate presented methods permanentlyTo extend its reasonable applicability forlarger and larger class of DES able to be modelled by Petri netsTo find new methods, procedures and toolsfor DEDS modelling, analyzing and control