Synthesis of Optimal Strategies for Differential Games by Neural Networks View Full Text


Ontology type: schema:Chapter     


Chapter Info

DATE

1995

AUTHORS

H. J. Pesch , I. Gabler , S. Miesbach , M. H. Breitner

ABSTRACT

The paper deals with the numerical approximation of optimal strategies for two-person zero-sum differential games of pursuit-evasion type by neural networks. Thereby, the feedback strategies can be computed in real-time after the training of appropriate neural networks. For this purpose, sufficiently many optimal trajectories and their associated open-loop representations of the optimal feedback strategies must be computed, to provide data for training and cross-validation of the neural networks. All the precomputations can be carried through in a highly parallel way. This approach turns out to be applicable for differential games of more general type.The method is demonstrated for a modified cornered rat game where a pursuing cat and an evading rat, both moving in simple motion, are constrained to a rectangular arena. Two holes in the walls surrounding the arena enable the rat to evade once and for all, if the rat is not too far from these holes. The optimal trajectories in the escape zone can be computed analytically. In the capture zone, a game of degree is employed with terminal time as payoff. To compute optimal trajectories for this secondary game, the time evolution of the survival region for the rat is determined via a sequence of discretized games.The combination of these methods permits the computation of more than a thousand trajectories leading to some ten thousand sample patterns which relate the state variables to the values of the optimal strategies. These data exhibit characteristic properties of the optimal strategies. It is shown that these properties can be extracted from the data by use of neural networks. By means of the trained networks, about 200 trajectories arc finally simulated. The pursuer as well as the evader acts according to the controls proposed by the neural networks. Despite the simple structure of the neural networks used in this study, the strategics based upon them show a reasonable, close to optimal performance in a large variety of simulations of the pursuit-evasion game under consideration. More... »

PAGES

111-141

Identifiers

URI

http://scigraph.springernature.com/pub.10.1007/978-1-4612-4274-1_6

DOI

http://dx.doi.org/10.1007/978-1-4612-4274-1_6

DIMENSIONS

https://app.dimensions.ai/details/publication/pub.1047808043


Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
Incoming Citations Browse incoming citations for this publication using opencitations.net

JSON-LD is the canonical representation for SciGraph data.

TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

[
  {
    "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
    "about": [
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/08", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Information and Computing Sciences", 
        "type": "DefinedTerm"
      }, 
      {
        "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/0801", 
        "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
        "name": "Artificial Intelligence and Image Processing", 
        "type": "DefinedTerm"
      }
    ], 
    "author": [
      {
        "affiliation": {
          "alternateName": "Department of Mathematics, Munich University of Technology, D-80290, Munich, Germany", 
          "id": "http://www.grid.ac/institutes/grid.6936.a", 
          "name": [
            "Department of Mathematics, Munich University of Technology, D-80290, Munich, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Pesch", 
        "givenName": "H. J.", 
        "id": "sg:person.014543723551.22", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014543723551.22"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Bayerische Hypotheken- und Wechsel-Bank AG, D-80278, Munich, Germany", 
          "id": "http://www.grid.ac/institutes/None", 
          "name": [
            "Bayerische Hypotheken- und Wechsel-Bank AG, D-80278, Munich, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Gabler", 
        "givenName": "I.", 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Corporate Research and Development, ZFE ST SN 41, Siemens AG, Munich, D-81730, Munich, Germany", 
          "id": "http://www.grid.ac/institutes/grid.5406.7", 
          "name": [
            "Corporate Research and Development, ZFE ST SN 41, Siemens AG, Munich, D-81730, Munich, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Miesbach", 
        "givenName": "S.", 
        "id": "sg:person.014243476305.41", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014243476305.41"
        ], 
        "type": "Person"
      }, 
      {
        "affiliation": {
          "alternateName": "Department of Mathematics, Munich University of Technology, D-80290, Munich, Germany", 
          "id": "http://www.grid.ac/institutes/grid.6936.a", 
          "name": [
            "Department of Mathematics, Munich University of Technology, D-80290, Munich, Germany"
          ], 
          "type": "Organization"
        }, 
        "familyName": "Breitner", 
        "givenName": "M. H.", 
        "id": "sg:person.010416245235.49", 
        "sameAs": [
          "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.010416245235.49"
        ], 
        "type": "Person"
      }
    ], 
    "datePublished": "1995", 
    "datePublishedReg": "1995-01-01", 
    "description": "The paper deals with the numerical approximation of optimal strategies for two-person zero-sum differential games of pursuit-evasion type by neural networks. Thereby, the feedback strategies can be computed in real-time after the training of appropriate neural networks. For this purpose, sufficiently many optimal trajectories and their associated open-loop representations of the optimal feedback strategies must be computed, to provide data for training and cross-validation of the neural networks. All the precomputations can be carried through in a highly parallel way. This approach turns out to be applicable for differential games of more general type.The method is demonstrated for a modified cornered rat game where a pursuing cat and an evading rat, both moving in simple motion, are constrained to a rectangular arena. Two holes in the walls surrounding the arena enable the rat to evade once and for all, if the rat is not too far from these holes. The optimal trajectories in the escape zone can be computed analytically. In the capture zone, a game of degree is employed with terminal time as payoff. To compute optimal trajectories for this secondary game, the time evolution of the survival region for the rat is determined via a sequence of discretized games.The combination of these methods permits the computation of more than a thousand trajectories leading to some ten thousand sample patterns which relate the state variables to the values of the optimal strategies. These data exhibit characteristic properties of the optimal strategies. It is shown that these properties can be extracted from the data by use of neural networks. By means of the trained networks, about 200 trajectories arc finally simulated. The pursuer as well as the evader acts according to the controls proposed by the neural networks. Despite the simple structure of the neural networks used in this study, the strategics based upon them show a reasonable, close to optimal performance in a large variety of simulations of the pursuit-evasion game under consideration.", 
    "editor": [
      {
        "familyName": "Olsder", 
        "givenName": "Geert Jan", 
        "type": "Person"
      }
    ], 
    "genre": "chapter", 
    "id": "sg:pub.10.1007/978-1-4612-4274-1_6", 
    "isAccessibleForFree": false, 
    "isPartOf": {
      "isbn": [
        "978-1-4612-8719-3", 
        "978-1-4612-4274-1"
      ], 
      "name": "New Trends in Dynamic Games and Applications", 
      "type": "Book"
    }, 
    "keywords": [
      "neural network", 
      "optimal trajectories", 
      "appropriate neural network", 
      "open-loop representations", 
      "pursuit-evasion game", 
      "parallel way", 
      "optimal strategy", 
      "network", 
      "sample patterns", 
      "feedback strategy", 
      "zero-sum differential game", 
      "optimal performance", 
      "optimal feedback strategy", 
      "simple motion", 
      "game", 
      "differential game", 
      "pursuit-evasion type", 
      "precomputation", 
      "large variety", 
      "two-person", 
      "more general types", 
      "trajectories", 
      "computation", 
      "training", 
      "evader", 
      "simple structure", 
      "game of degree", 
      "survival region", 
      "pursuer", 
      "representation", 
      "numerical approximation", 
      "data", 
      "terminal time", 
      "state variables", 
      "method", 
      "performance", 
      "time evolution", 
      "strategies", 
      "Strategic", 
      "simulations", 
      "general type", 
      "capture zone", 
      "way", 
      "escape zone", 
      "payoffs", 
      "arena", 
      "characteristic properties", 
      "approximation", 
      "variety", 
      "motion", 
      "time", 
      "use", 
      "types", 
      "sequence", 
      "purpose", 
      "holes", 
      "means", 
      "consideration", 
      "rectangular arena", 
      "properties", 
      "control", 
      "combination", 
      "patterns", 
      "structure", 
      "evolution", 
      "variables", 
      "approach", 
      "degree", 
      "values", 
      "zone", 
      "region", 
      "study", 
      "wall", 
      "synthesis", 
      "cats", 
      "paper", 
      "rats"
    ], 
    "name": "Synthesis of Optimal Strategies for Differential Games by Neural Networks", 
    "pagination": "111-141", 
    "productId": [
      {
        "name": "dimensions_id", 
        "type": "PropertyValue", 
        "value": [
          "pub.1047808043"
        ]
      }, 
      {
        "name": "doi", 
        "type": "PropertyValue", 
        "value": [
          "10.1007/978-1-4612-4274-1_6"
        ]
      }
    ], 
    "publisher": {
      "name": "Springer Nature", 
      "type": "Organisation"
    }, 
    "sameAs": [
      "https://doi.org/10.1007/978-1-4612-4274-1_6", 
      "https://app.dimensions.ai/details/publication/pub.1047808043"
    ], 
    "sdDataset": "chapters", 
    "sdDatePublished": "2022-12-01T06:47", 
    "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
    "sdPublisher": {
      "name": "Springer Nature - SN SciGraph project", 
      "type": "Organization"
    }, 
    "sdSource": "s3://com-springernature-scigraph/baseset/20221201/entities/gbq_results/chapter/chapter_166.jsonl", 
    "type": "Chapter", 
    "url": "https://doi.org/10.1007/978-1-4612-4274-1_6"
  }
]
 

Download the RDF metadata as:  json-ld nt turtle xml License info

HOW TO GET THIS DATA PROGRAMMATICALLY:

JSON-LD is a popular format for linked data which is fully compatible with JSON.

curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/978-1-4612-4274-1_6'

N-Triples is a line-based linked data format ideal for batch operations.

curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/978-1-4612-4274-1_6'

Turtle is a human-readable linked data format.

curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/978-1-4612-4274-1_6'

RDF/XML is a standard XML format for linked data.

curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/978-1-4612-4274-1_6'


 

This table displays all metadata directly associated to this object as RDF triples.

162 TRIPLES      22 PREDICATES      102 URIs      95 LITERALS      7 BLANK NODES

Subject Predicate Object
1 sg:pub.10.1007/978-1-4612-4274-1_6 schema:about anzsrc-for:08
2 anzsrc-for:0801
3 schema:author N9e9f8d7d5b1048c499bd3e488317a9cf
4 schema:datePublished 1995
5 schema:datePublishedReg 1995-01-01
6 schema:description The paper deals with the numerical approximation of optimal strategies for two-person zero-sum differential games of pursuit-evasion type by neural networks. Thereby, the feedback strategies can be computed in real-time after the training of appropriate neural networks. For this purpose, sufficiently many optimal trajectories and their associated open-loop representations of the optimal feedback strategies must be computed, to provide data for training and cross-validation of the neural networks. All the precomputations can be carried through in a highly parallel way. This approach turns out to be applicable for differential games of more general type.The method is demonstrated for a modified cornered rat game where a pursuing cat and an evading rat, both moving in simple motion, are constrained to a rectangular arena. Two holes in the walls surrounding the arena enable the rat to evade once and for all, if the rat is not too far from these holes. The optimal trajectories in the escape zone can be computed analytically. In the capture zone, a game of degree is employed with terminal time as payoff. To compute optimal trajectories for this secondary game, the time evolution of the survival region for the rat is determined via a sequence of discretized games.The combination of these methods permits the computation of more than a thousand trajectories leading to some ten thousand sample patterns which relate the state variables to the values of the optimal strategies. These data exhibit characteristic properties of the optimal strategies. It is shown that these properties can be extracted from the data by use of neural networks. By means of the trained networks, about 200 trajectories arc finally simulated. The pursuer as well as the evader acts according to the controls proposed by the neural networks. Despite the simple structure of the neural networks used in this study, the strategics based upon them show a reasonable, close to optimal performance in a large variety of simulations of the pursuit-evasion game under consideration.
7 schema:editor N55bcabae7aba422482c3f431bf8cc22a
8 schema:genre chapter
9 schema:isAccessibleForFree false
10 schema:isPartOf N685a5da00d5f4d87af2462b408de919e
11 schema:keywords Strategic
12 approach
13 appropriate neural network
14 approximation
15 arena
16 capture zone
17 cats
18 characteristic properties
19 combination
20 computation
21 consideration
22 control
23 data
24 degree
25 differential game
26 escape zone
27 evader
28 evolution
29 feedback strategy
30 game
31 game of degree
32 general type
33 holes
34 large variety
35 means
36 method
37 more general types
38 motion
39 network
40 neural network
41 numerical approximation
42 open-loop representations
43 optimal feedback strategy
44 optimal performance
45 optimal strategy
46 optimal trajectories
47 paper
48 parallel way
49 patterns
50 payoffs
51 performance
52 precomputation
53 properties
54 purpose
55 pursuer
56 pursuit-evasion game
57 pursuit-evasion type
58 rats
59 rectangular arena
60 region
61 representation
62 sample patterns
63 sequence
64 simple motion
65 simple structure
66 simulations
67 state variables
68 strategies
69 structure
70 study
71 survival region
72 synthesis
73 terminal time
74 time
75 time evolution
76 training
77 trajectories
78 two-person
79 types
80 use
81 values
82 variables
83 variety
84 wall
85 way
86 zero-sum differential game
87 zone
88 schema:name Synthesis of Optimal Strategies for Differential Games by Neural Networks
89 schema:pagination 111-141
90 schema:productId N8200b349a97a4496a68eaa742bb02c22
91 Nb1a2111e8a444273b2854eca8c73c38f
92 schema:publisher N1af7a4cad55f44559601b09a816ec686
93 schema:sameAs https://app.dimensions.ai/details/publication/pub.1047808043
94 https://doi.org/10.1007/978-1-4612-4274-1_6
95 schema:sdDatePublished 2022-12-01T06:47
96 schema:sdLicense https://scigraph.springernature.com/explorer/license/
97 schema:sdPublisher N52c71f17382542118473739c6e1760aa
98 schema:url https://doi.org/10.1007/978-1-4612-4274-1_6
99 sgo:license sg:explorer/license/
100 sgo:sdDataset chapters
101 rdf:type schema:Chapter
102 N0c0c41cbd28748a6b5cc41f7b644eed5 rdf:first N837c50d9d84842ffb06567a556db0841
103 rdf:rest Nf5ff810e56144d4ea3a4ecb08faa1c5a
104 N1af7a4cad55f44559601b09a816ec686 schema:name Springer Nature
105 rdf:type schema:Organisation
106 N4505415d52b242b889afb34118725cdf rdf:first sg:person.010416245235.49
107 rdf:rest rdf:nil
108 N52c71f17382542118473739c6e1760aa schema:name Springer Nature - SN SciGraph project
109 rdf:type schema:Organization
110 N55bcabae7aba422482c3f431bf8cc22a rdf:first N8ca925acfdd14387ab5ade8de5756c50
111 rdf:rest rdf:nil
112 N685a5da00d5f4d87af2462b408de919e schema:isbn 978-1-4612-4274-1
113 978-1-4612-8719-3
114 schema:name New Trends in Dynamic Games and Applications
115 rdf:type schema:Book
116 N8200b349a97a4496a68eaa742bb02c22 schema:name doi
117 schema:value 10.1007/978-1-4612-4274-1_6
118 rdf:type schema:PropertyValue
119 N837c50d9d84842ffb06567a556db0841 schema:affiliation grid-institutes:None
120 schema:familyName Gabler
121 schema:givenName I.
122 rdf:type schema:Person
123 N8ca925acfdd14387ab5ade8de5756c50 schema:familyName Olsder
124 schema:givenName Geert Jan
125 rdf:type schema:Person
126 N9e9f8d7d5b1048c499bd3e488317a9cf rdf:first sg:person.014543723551.22
127 rdf:rest N0c0c41cbd28748a6b5cc41f7b644eed5
128 Nb1a2111e8a444273b2854eca8c73c38f schema:name dimensions_id
129 schema:value pub.1047808043
130 rdf:type schema:PropertyValue
131 Nf5ff810e56144d4ea3a4ecb08faa1c5a rdf:first sg:person.014243476305.41
132 rdf:rest N4505415d52b242b889afb34118725cdf
133 anzsrc-for:08 schema:inDefinedTermSet anzsrc-for:
134 schema:name Information and Computing Sciences
135 rdf:type schema:DefinedTerm
136 anzsrc-for:0801 schema:inDefinedTermSet anzsrc-for:
137 schema:name Artificial Intelligence and Image Processing
138 rdf:type schema:DefinedTerm
139 sg:person.010416245235.49 schema:affiliation grid-institutes:grid.6936.a
140 schema:familyName Breitner
141 schema:givenName M. H.
142 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.010416245235.49
143 rdf:type schema:Person
144 sg:person.014243476305.41 schema:affiliation grid-institutes:grid.5406.7
145 schema:familyName Miesbach
146 schema:givenName S.
147 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014243476305.41
148 rdf:type schema:Person
149 sg:person.014543723551.22 schema:affiliation grid-institutes:grid.6936.a
150 schema:familyName Pesch
151 schema:givenName H. J.
152 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.014543723551.22
153 rdf:type schema:Person
154 grid-institutes:None schema:alternateName Bayerische Hypotheken- und Wechsel-Bank AG, D-80278, Munich, Germany
155 schema:name Bayerische Hypotheken- und Wechsel-Bank AG, D-80278, Munich, Germany
156 rdf:type schema:Organization
157 grid-institutes:grid.5406.7 schema:alternateName Corporate Research and Development, ZFE ST SN 41, Siemens AG, Munich, D-81730, Munich, Germany
158 schema:name Corporate Research and Development, ZFE ST SN 41, Siemens AG, Munich, D-81730, Munich, Germany
159 rdf:type schema:Organization
160 grid-institutes:grid.6936.a schema:alternateName Department of Mathematics, Munich University of Technology, D-80290, Munich, Germany
161 schema:name Department of Mathematics, Munich University of Technology, D-80290, Munich, Germany
162 rdf:type schema:Organization
 




Preview window. Press ESC to close (or click here)


...