Control of traffic systems in buildings : applications of modern supervisory and optimal control

Control of Traffic Systems in Buildings, focusing on elevator groups, presents the state of the art in their analysis and control. It covers the theory and design of passenger and cargo traffic systems with actual operational examples and topics of current interest such as: noisy, on-line and algorithmic optimization; identification and tracking of loads; integration of control with security and management; medium- and wide-area networked control; deployment and testing of transport systems.

Workers in elevator control have pioneered the development of many modern control systems for use in all sorts of traffic and scheduled systems so this exposition of recent work in in-building transport control will be of interest to researchers and engineers in many areas of control, particularly in optimal or supervisory control, urban transportation systems and intelligent transport systems.

Author Notes

Sandor Markon has 25 years experience as a designer, researcher and manager dealing with control systems with major elevator manufacturers. He is also active in publishing papers and abstracts in related fields. He has been involved in the installation of four generations of elevator systems in buildings world-wide. The current volume would gathers that knowledge together with that of other researchers currently scattered in short articles, patent disclosures, etc., into one place.

Part I Transportation Systems
1 Introduction	p. 3
2 Passenger Transportation Systems	p. 7
2.1 Elevators	p. 9
2.1.1 Construction and Operation	p. 10
2.1.2 Safety	p. 10
2.1.3 Modern Technology	p. 10
2.1.4 Control	p. 10
2.2 Other Passenger Transportation Equipment	p. 11
2.2.1 Escalators	p. 11
2.2.2 Moving Walkways	p. 13
2.2.3 Horizontal Elevators	p. 13
3 Cargo Transportation Systems	p. 15
3.1 Freight Elevators	p. 15
3.2 Conveyors	p. 15
3.3 Automated Guided Vehicles	p. 17
3.4 Stacker Cranes	p. 18
4 External Connections and Related Systems	p. 19
4.1 External Connections	p. 19
4.1.1 Pedestrian Connections	p. 19
4.1.2 Freight Connections	p. 19
4.2 Related Systems	p. 19
4.2.1 Factory Automation	p. 20
4.2.2 Warehouse Automation	p. 20
4.2.3 Hospital Automation	p. 20
Part II Modeling and Simulation
5 General Modeling Concepts	p. 23
5.1 Components and Topology	p. 23
5.1.1 Vehicles	p. 23
5.1.2 Guideways	p. 24
5.1.3 Signal Systems	p. 26
5.1.4 Zones and Banks	p. 28
5.1.5 Nodes and Links	p. 29
5.2 Human-machine Interaction and Control Objectives	p. 30
5.2.1 Modeling of the Traffic	p. 30
5.2.2 Human-machine Interface of Elevators	p. 31
5.2.3 Human-machine Interface of Escalators and Other Equipment	p. 32
5.2.4 Control Objectives	p. 32
6 Queuing Models	p. 33
6.1 General Overview of Queuing Models	p. 33
6.2 Queuing Models for Elevator Systems	p. 34
6.2.1 The Simplest Case: M/M/1 Model	p. 34
6.2.2 A More General Model: M/G/1	p. 36
7 Modeling Techniques for Discrete Event Systems	p. 39
7.1 Field Studies	p. 39
7.2 Monte-Carlo Modeling	p. 41
7.2.1 Simulation Techniques	p. 41
7.2.2 Modeling by ESM-based Methodology	p. 41
7.3 The ESM Framework for Simulations	p. 43
7.3.1 The ESM Model for Discrete Event Simulation	p. 43
7.3.2 Communication Between ESMs	p. 45
7.3.3 Tools for Defining the ESM Model	p. 47
7.3.4 Implementation of the Simulation Program	p. 48
7.4 Modeling Cooperating Elevators and AGVs by the ESM Methodology	p. 50
7.4.1 Traffic Survey as the Starting Point for Simulations	p. 51
7.4.2 A Simplified Model of the Traffic in the Building	p. 52
8 Scheduling Models with Transportation	p. 55
8.1 Jobshop Scheduling Problems	p. 55
8.2 Classification of Jobshop Scheduling Problems	p. 61
8.3 Computational Complexity and Optimization Methods for JSP	p. 62
8.4 Robotic Cell Scheduling Problems	p. 64
Part III Intelligent Control Methods for Transportation Systems
9 Analytical and Heuristic Control of Transportation Systems	p. 69
9.1 Evolution of Control Methods	p. 69
9.2 Analytical Approaches	p. 70
9.3 Heuristic Rules	p. 71
9.3.1 Algorithmic Control	p. 72
9.3.2 Fuzzy AI Group Control	p. 73
9.4 Early Approaches to Optimal Control	p. 74
10 Adaptive Control by Neural Networks and Reinforcement Learning	p. 79
10.1 Information Processing by Neural Networks	p. 79
10.2 Multilayer Perceptrons	p. 80
10.2.1 Model of the Processing Units	p. 80
10.2.2 Structure and Operation of the Multilayer Perceptron	p. 80
10.2.3 Expressive Power of the MLP	p. 82
10.3 Learning as an Optimization Problem	p. 83
10.3.1 Nonlinear Optimization by the Gradient Method	p. 84
10.3.2 Derivation of the Learning Rule	p. 85
10.3.3 Hints for the Implementation and Use of the BP Method	p. 87
10.3.4 Using More Refined Optimization Methods	p. 89
10.4 Learning and Generalization by MLPs	p. 91
10.4.1 Learning and Generalization	p. 91
10.4.2 Generalization in the Case of MLPs	p. 91
10.4.3 Testing MLPs	p. 91
10.4.4 Learning by Direct Optimization	p. 92
10.4.5 Forward-Backward Modeling	p. 92
10.4.6 Learning with Powell's Conjugate Direction Method	p. 93
10.4.7 Learning by Genetic Algorithms	p. 93
10.5 Reinforcement Learning	p. 94
10.5.1 Markov Decision Processes	p. 94
10.5.2 Dynamic Programming (DP)	p. 96
10.5.3 The Value Iteration Method	p. 97
10.5.4 Q-learning	p. 98
11 Genetic Algorithms for Control-system Optimization	p. 103
11.1 Stochastic Approach to Optimization	p. 103
11.2 Genetic Algorithm	p. 104
11.2.1 Combinatorial Optimization with GA	p. 105
11.2.2 Nonlinear Optimization with GA	p. 107
11.2.3 GA as the Evolution of Distributions	p. 108
11.2.4 GA and Estimation of Distributions Algorithms	p. 110
11.3 Optimization of Uncertain Fitness Functions by Genetic Algorithms	p. 111
11.3.1 Introduction to GA for Optimization with Uncertainty	p. 111
11.3.2 Optimization of Noisy Fitness Functions	p. 112
11.3.3 Adaptation to Changing Environment	p. 112
11.3.4 Discussion from the Application Side	p. 113
11.3.5 Approach to Uncertain Optimization by GA	p. 114
11.3.6 GA for Optimizing a Fitness Function with Noise	p. 115
11.3.7 GA for Varying Environments	p. 116
11.3.8 MFEGA and an Example of its Application	p. 118
12 Control System Optimization by ES and PSO	p. 121
12.1 Evolution Strategies	p. 121
12.1.1 Framework of Evolution Strategies	p. 121
12.1.2 Algorithm Designs for Evolutionary Algorithms	p. 121
12.2 Optimization of Noisy Fitness with Evolution Strategies	p. 128
12.2.1 Ways to Cope with Uncertainty	p. 129
12.2.2 Optimal Computing Budget Allocation	p. 131
12.2.3 Threshold Selection	p. 132
12.3 Particle Swarm Optimization	p. 137
12.3.1 Framework of Particle Swarm Optimization	p. 137
12.3.2 PSO and Noisy Optimization Problem	p. 139
12.4 Summary	p. 141
13 Intelligent Control by Combinatorial Optimization	p. 143
13.1 Branch-and-Bound Search	p. 143
13.2 Tabu Search	p. 145
13.2.1 Definition of the Problem	p. 145
13.2.2 Local Search	p. 145
13.2.3 Basic Structure of Tabu Search	p. 147
Part IV Topics in Modern Control for Transportation Systems
14 The S-ring: a Transportation System Model for Benchmarking
14.1 The Kac Ring	p. 151
14.2 Definition of the S-ring Model	p. 151
14.3 Control of the S-ring	p. 153
14.3.1 Representations of the Policy	p. 156
14.3.2 Policy Examples	p. 156
14.3.3 Extensions	p. 157
14.4 A Prototype S-ring	p. 158
14.5 Solution by Dynamic Programming	p. 158
14.5.1 Formulation	p. 158
14.5.2 Solution	p. 159
14.6 Solution by Numerical Methods	p. 159
14.6.1 Kiefer-Wolfowitz Stochastic Approximation	p. 160
14.6.2 Q-learning and Evolutionary Strategies	p. 160
14.6.3 Results of the Optimization Experiments	p. 161
14.7 Conclusions	p. 161
15 Elevator Group Control by NN and Stochastic Approximation	p. 163
15.1 The Elevator Group Control as an Optimal Control Problem	p. 164
15.2 Elevator Group Control by Neural Networks	p. 165
15.2.1 State Representation for Elevator Group Control	p. 166
15.3 Neurocontroller for Group Control	p. 169
15.3.1 Structure of the Neurocontroller for Elevator Group Control	p. 171
15.3.2 Initial Training of the Neurocontroller	p. 174
15.4 Adaptive Optimal Control by the Stochastic Approximation	p. 177
15.4.1 Outline of the Basic Adaptation Process	p. 177
15.4.2 Sensitivity of the Controller Network	p. 179
15.4.3 Simulation Results for Adaptive Optimal Group Control	p. 182
15.5 Conclusions	p. 186
16 Optimal Control by Evolution Strategies and PSO	p. 187
16.1 Sequential Parameter Optimization	p. 188
16.1.1 SPO as a Learning Tool	p. 188
16.1.2 Tuning	p. 190
16.1.3 Stochastic Process Models as Extensions of Classical Regression Models	p. 191
16.1.4 Space-filling Designs	p. 195
16.2 The S-ring Model as a Test Generator	p. 195
16.3 Experimental Results for the S-ring Model	p. 198
16.3.1 Evolution Strategies	p. 198
16.3.2 Particle Swarm Optimization on the S-ring Model	p. 203
16.4 Classical Algorithms on the S-ring Model	p. 208
16.5 Criteria for Choosing an Optimization Algorithm	p. 209
17 On Adaptive Cooperation of AGVs and Elevators	p. 211
17.1 Introduction	p. 211
17.2 Material Handling System for High-rise Buildings	p. 212
17.3 Contract Net Protocol	p. 213
17.4 Intrabuilding Traffic Simulator	p. 214
17.4.1 Outline of the Simulator	p. 214
17.4.2 Performance Index of Control	p. 214
17.5 Cooperation based on Estimated Processing Time	p. 216
17.5.1 Control Using Minimal Processing Time for Bidding	p. 216
17.5.2 Estimation of Processing Time by a Neural Network	p. 216
17.5.3 Numerical Example	p. 217
17.6 Optimization of Performance	p. 218
17.6.1 Bidding Function to be Optimized	p. 218
17.6.2 Application of Genetic Algorithm	p. 218
17.6.3 Numerical Example	p. 219
17.7 Conclusion	p. 219
18 Optimal Control of Multicar Elevator Systems by Genetic Algorithms	p. 221
18.1 Introduction	p. 221
18.2 Multicar Elevator Systems and Controller Optimization	p. 222
18.2.1 Multicar Elevator Systems	p. 222
18.2.2 Controllers for MCE	p. 223
18.2.3 Discrete Event Simulation of MCE	p. 223
18.2.4 Simulation-based Optimization	p. 224
18.2.5 Problems in Optimization	p. 225
18.2.6 Acceleration of Computation	p. 225
18.2.7 Re-examination of Configuration of Simulation	p. 226
18.3 A Genetic Algorithm for Noisy Fitness Function	p. 226
18.4 Comparison of GAs for Noisy Fitness	p. 227
18.4.1 Setup of Experiments	p. 227
18.4.2 Results of Experiment	p. 228
18.5 Examination of Control Strategy	p. 230
18.5.1 Examination of Zone Boundary	p. 230
18.5.2 Effect of Weight Extension	p. 230
18.6 Conclusion	p. 232
19 Analysis and Optimization for Automated Vehicle Routing	p. 235
19.1 Introduction	p. 235
19.2 Basic Assumptions and Basic Analysis	p. 236
19.2.1 Parallel and Bottleneck-Free PCVRS	p. 236
19.2.2 Interferences and Steady State	p. 237
19.2.3 One Lap Behind Interference	p. 239
19.2.4 Throughput and Mean Interference Time	p. 240
19.3 Two Basic Vehicle Routings	p. 241
19.3.1 Random Rule	p. 242
19.3.2 Order Rule	p. 242
19.4 Optimal Vehicle Rules	p. 244
19.4.1 Exchange-Order Rule	p. 244
19.4.2 Dynamic Order Rule	p. 247
19.5 Numerical Simulation	p. 247
19.6 Concluding Remarks	p. 249
20 Tabu-based Optimization for Input/Output Scheduling	p. 251
20.1 Introduction	p. 251
20.2 Optimal Input/Output Scheduling Problem	p. 251
20.3 Computational Complexity	p. 252
20.4 Approximation Algorithm	p. 253
20.5 Numerical Experiment	p. 255
20.6 Concluding Remarks	p. 255
Program Listings	p. 257
References	p. 261
Index	p. 275

Available:*

On Order

Summary

Summary

Author Notes

Table of Contents