Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optim... more Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optimization problems are investigated in this paper. In the problems studied here, the fitness function changes during the search carried out by the GA. In the GA investigated, each gene is associated with an independent mutation probability. The knowledge obtained during the evolution is utilized to update the mutation probabilities. If the modification of a set of genes is useful when the problem changes, the mutation probabilities of these genes are increased. In this way, the search in the solution space is concentrated into regions associated with the genes with higher mutation probabilities. The class of non-stationary problems where this GA can be interesting and its limitations are investigated.
In this paper a genetic algorithm is proposed where the worst individual and individuals with ind... more In this paper a genetic algorithm is proposed where the worst individual and individuals with indices close to its index are replaced in every generation by randomly generated individuals for dynamic optimization problems. In the proposed genetic algorithm, the replacement of an individual can affect other individuals in a chain reaction. The new individuals are preserved in a subpopulation which is defined by the number of individuals created in the current chain reaction. If the values of fitness are similar, as is the case with small diversity, one single replacement can affect a large number of individuals in the population. This simple approach can take the system to a self-organizing behavior, which can be useful to control the diversity level of the population and hence allows the genetic algorithm to escape from local optima once the problem changes due to the dynamics.
Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optim... more Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optimization problems are investigated in this paper. In the problems studied here, the fitness function changes during the search carried out by the GA. In the GA investigated, each gene is associated with an independent mutation probability. The knowledge obtained during the evolution is utilized to update the mutation probabilities. If the modification of a set of genes is useful when the problem changes, the mutation probabilities of these genes are increased. In this way, the search in the solution space is concentrated into regions associated with the genes with higher mutation probabilities. The class of non-stationary problems where this GA can be interesting and its limitations are investigated.
In this article, the authors investigate the application of genetic algorithms (GAs) with gene de... more In this article, the authors investigate the application of genetic algorithms (GAs) with gene dependent mutation probability to the training of artificial neural networks (ANNs) in non-stationary problems (NSPs). In the problems studied, the function mapped by an ANN changes during the search carried out by the GA. In the GA proposed, each gene is associated with an independent mutation probability. The knowledge obtained during the evolution is used to update the mutation probabilities. If the modification of a set of genes is useful when the problem changes its profile, the mutation probabilities of these genes are increased. As a result, the search is concentrated into regions associated with genes presenting higher mutation probabilities.
The use of evolutionary programming algorithms with self-adaptation of the mutation distribution ... more The use of evolutionary programming algorithms with self-adaptation of the mutation distribution for dynamic optimization problems is investigated in this paper. In the proposed method, the q-Gaussian distribution is employed to generate new candidate solutions by mutation. A real parameter q, which defines the shape of the distribution, is encoded in the chromosome of individuals and is allowed to evolve. Algorithms with self-adapted mutation generated from isotropic and anisotropic distributions are presented. In the experimental study, the q-Gaussian mutation is compared to Gaussian and Cauchy mutation on three dynamic optimization problems.
In this paper we consider the use of associative search and adaptive critic elements and artifici... more In this paper we consider the use of associative search and adaptive critic elements and artificial neural network for control of nonlinear and unstable plants. The reinforcement learning schemes we propose are used in the design of different controllers. An example of a magnetic suspension system is presented to illustrate the effectiveness of these controllers. We also include results of a linear optimal controller
Usually, fault detection and isolation schemes for robotic manipulators use the system mathematic... more Usually, fault detection and isolation schemes for robotic manipulators use the system mathematical model to generate the residual vector. However, modeling errors could obscure the faults and could be a false alarm source. In this paper a multilayer perceptron trained with backpropagation algorithm is employed to reproduce the robot input/output behavior generating the residual vector. Then, a radial basis function network is utilized to classify the residual vector generating the fault isolation. Three different algorithms have been employed to train this network. The first employs subset selection to choose the radial units from the training patterns. The second utilizes regularization to reduce the variance of the model. The third algorithm also uses regularization but, instead of one penalty term, each radial unit has an individual penalty term. Simulations employing a two-link manipulator are showed demonstrating that the system can detect and isolate correctly faults that occur in nontrained trajectories
In recent years, several approaches have been developed for genetic algorithms to enhance their p... more In recent years, several approaches have been developed for genetic algorithms to enhance their performance in dynamic environments. Among these approaches, one kind of methods is to adapt genetic operators in order for genetic algorithms to adapt to a new environment. This paper investigates the effect of the selection pressure on the performance of genetic algorithms in dynamic environments. A hyper-selection scheme is proposed for genetic algorithms, where the selection pressure is temporarily raised whenever the environment changes. The hyper-selection scheme can be combined with other approaches for genetic algorithms in dynamic environments. Experiments are carried out to investigate the effect of different selection pressures on the performance of genetic algorithms in dynamic environments and to investigate the effect of the hyper-selection scheme on the performance of genetic algorithms in combination with several other schemes in dynamic environments. The experimental results indicate that the effect of the hyper-selection scheme depends on the problem under consideration and other schemes combined in genetic algorithms.
This paper proposes a self-adaptation method to control not only the mutation strength parameter,... more This paper proposes a self-adaptation method to control not only the mutation strength parameter, but also the mutation distribution for evolutionary algorithms. For this purpose, the isotropic g-Gaussian distribution is employed in the mutation operator. The g-Gaussian distribution allows to control the shape of the distribution by setting a real parameter g and can reproduce either finite second moment distributions or infinite second moment distributions. In the proposed method, the real parameter q of the g-Gaussian distribution is encoded in the chromosome of an individual and is allowed to evolve. An evolutionary programming algorithm with the proposed idea is presented. Experiments were carried out to study the performance of the proposed algorithm.
Addressing dynamic optimization problems has attracted a growing interest from the evolutionary a... more Addressing dynamic optimization problems has attracted a growing interest from the evolutionary algorithm community in recent years due to its importance in the applications of evolutionary algorithms in real world problems. In order to study evolutionary algorithms in dynamic environments, one important work is to develop benchmark dynamic environments. This paper proposes two continuous dynamic problem generators. Both generators use linear transformation to move individuals, which preserves the distance among individuals. In the first generator, the linear transformation of individuals is equivalent to change the direction of some axes of the search space while in the second one it is obtained by successive rotations in different planes. Preliminary experiments were carried out to study the performance of some standard genetic algorithms in continuous dynamic environments created by the proposed generators.
This paper proposes a genetic algorithm (GA) with random immigrants for dynamic optimization prob... more This paper proposes a genetic algorithm (GA) with random immigrants for dynamic optimization problems where the worst individual and its neighbours are replaced every generation. In this GA, the individuals interact with each other and, when their fitness is close, as in the case where the diversity level is low, one single replacement can affect a large number of individuals. This simple approach can take the system to a kind of self-organization behavior, known as self-organized criticality (SOC), which is useful to maintain the diversity of the population in dynamic environments and hence allows the GA to escape from local optima when the problem changes. The experimental results show that the proposed GA presents the phenomenon of SOC.
The problem of fault tolerance in cooperative manipulators rigidly connected to a solid object is... more The problem of fault tolerance in cooperative manipulators rigidly connected to a solid object is addressed in this paper. Four faults are considered: free-swinging joint faults, locked joint faults, incorrect measured joint position, and incorrect measured joint velocity. The faults are first detected by a, fault detection and isolation system. Free-swinging and locked joint faults are isolated using artificial neural networks. The other faults are isolated based on the kinematic constraints imposed on the cooperative system. After the isolation of the faults, the control system is reconfigured. Control laws for the system with passive or locked joints are developed. Results of the fault tolerance system applied in simulations and in a real cooperative system are presented.
Faults in robotic manipulators can cause economic losses and serious damages. In the paper, two a... more Faults in robotic manipulators can cause economic losses and serious damages. In the paper, two artificial neural networks are employed to provide FDI to robotic manipulators. The first is a multilayer perceptron trained with backpropagation utilized to reproduce the dynamic of the manipulator and, so, generate the residual vector. The second is a radial basis function network employed to classify the residual vector and, thus, generate the fault isolation. As the system model is not employed, false alarms due to modeling errors are avoided. Two different algorithms are employed to train the last network. The first employs ridge regression (a regularization type) and the second uses forward selection (an algorithm for subset selection). Simulations in a two link manipulator evince that the FDI system can detect and isolate correctly faults that occur in nontrained trajectories
The problem of the control of cooperative manipulators with passive joints and are rigidly connec... more The problem of the control of cooperative manipulators with passive joints and are rigidly connected to a solid object is addressed in this paper. Passive joints can appear due to free-swinging joint failures or can be an intrinsic characteristic of the robots. A hybrid control of motion and squeeze force is proposed. For this purpose, a Jacobian matrix relating velocities in the actuated joints and load velocity is obtained based on the kinematic constraints of the cooperative system. Results of the control system applied in simulations and in real robots are presented.
Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optim... more Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optimization problems are investigated in this paper. In the problems studied here, the fitness function changes during the search carried out by the GA. In the GA investigated, each gene is associated with an independent mutation probability. The knowledge obtained during the evolution is utilized to update the mutation probabilities. If the modification of a set of genes is useful when the problem changes, the mutation probabilities of these genes are increased. In this way, the search in the solution space is concentrated into regions associated with the genes with higher mutation probabilities. The class of non-stationary problems where this GA can be interesting and its limitations are investigated.
In this paper a genetic algorithm is proposed where the worst individual and individuals with ind... more In this paper a genetic algorithm is proposed where the worst individual and individuals with indices close to its index are replaced in every generation by randomly generated individuals for dynamic optimization problems. In the proposed genetic algorithm, the replacement of an individual can affect other individuals in a chain reaction. The new individuals are preserved in a subpopulation which is defined by the number of individuals created in the current chain reaction. If the values of fitness are similar, as is the case with small diversity, one single replacement can affect a large number of individuals in the population. This simple approach can take the system to a self-organizing behavior, which can be useful to control the diversity level of the population and hence allows the genetic algorithm to escape from local optima once the problem changes due to the dynamics.
Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optim... more Genetic algorithms (GAs) with gene dependent mutation probability applied to non-stationary optimization problems are investigated in this paper. In the problems studied here, the fitness function changes during the search carried out by the GA. In the GA investigated, each gene is associated with an independent mutation probability. The knowledge obtained during the evolution is utilized to update the mutation probabilities. If the modification of a set of genes is useful when the problem changes, the mutation probabilities of these genes are increased. In this way, the search in the solution space is concentrated into regions associated with the genes with higher mutation probabilities. The class of non-stationary problems where this GA can be interesting and its limitations are investigated.
In this article, the authors investigate the application of genetic algorithms (GAs) with gene de... more In this article, the authors investigate the application of genetic algorithms (GAs) with gene dependent mutation probability to the training of artificial neural networks (ANNs) in non-stationary problems (NSPs). In the problems studied, the function mapped by an ANN changes during the search carried out by the GA. In the GA proposed, each gene is associated with an independent mutation probability. The knowledge obtained during the evolution is used to update the mutation probabilities. If the modification of a set of genes is useful when the problem changes its profile, the mutation probabilities of these genes are increased. As a result, the search is concentrated into regions associated with genes presenting higher mutation probabilities.
The use of evolutionary programming algorithms with self-adaptation of the mutation distribution ... more The use of evolutionary programming algorithms with self-adaptation of the mutation distribution for dynamic optimization problems is investigated in this paper. In the proposed method, the q-Gaussian distribution is employed to generate new candidate solutions by mutation. A real parameter q, which defines the shape of the distribution, is encoded in the chromosome of individuals and is allowed to evolve. Algorithms with self-adapted mutation generated from isotropic and anisotropic distributions are presented. In the experimental study, the q-Gaussian mutation is compared to Gaussian and Cauchy mutation on three dynamic optimization problems.
In this paper we consider the use of associative search and adaptive critic elements and artifici... more In this paper we consider the use of associative search and adaptive critic elements and artificial neural network for control of nonlinear and unstable plants. The reinforcement learning schemes we propose are used in the design of different controllers. An example of a magnetic suspension system is presented to illustrate the effectiveness of these controllers. We also include results of a linear optimal controller
Usually, fault detection and isolation schemes for robotic manipulators use the system mathematic... more Usually, fault detection and isolation schemes for robotic manipulators use the system mathematical model to generate the residual vector. However, modeling errors could obscure the faults and could be a false alarm source. In this paper a multilayer perceptron trained with backpropagation algorithm is employed to reproduce the robot input/output behavior generating the residual vector. Then, a radial basis function network is utilized to classify the residual vector generating the fault isolation. Three different algorithms have been employed to train this network. The first employs subset selection to choose the radial units from the training patterns. The second utilizes regularization to reduce the variance of the model. The third algorithm also uses regularization but, instead of one penalty term, each radial unit has an individual penalty term. Simulations employing a two-link manipulator are showed demonstrating that the system can detect and isolate correctly faults that occur in nontrained trajectories
In recent years, several approaches have been developed for genetic algorithms to enhance their p... more In recent years, several approaches have been developed for genetic algorithms to enhance their performance in dynamic environments. Among these approaches, one kind of methods is to adapt genetic operators in order for genetic algorithms to adapt to a new environment. This paper investigates the effect of the selection pressure on the performance of genetic algorithms in dynamic environments. A hyper-selection scheme is proposed for genetic algorithms, where the selection pressure is temporarily raised whenever the environment changes. The hyper-selection scheme can be combined with other approaches for genetic algorithms in dynamic environments. Experiments are carried out to investigate the effect of different selection pressures on the performance of genetic algorithms in dynamic environments and to investigate the effect of the hyper-selection scheme on the performance of genetic algorithms in combination with several other schemes in dynamic environments. The experimental results indicate that the effect of the hyper-selection scheme depends on the problem under consideration and other schemes combined in genetic algorithms.
This paper proposes a self-adaptation method to control not only the mutation strength parameter,... more This paper proposes a self-adaptation method to control not only the mutation strength parameter, but also the mutation distribution for evolutionary algorithms. For this purpose, the isotropic g-Gaussian distribution is employed in the mutation operator. The g-Gaussian distribution allows to control the shape of the distribution by setting a real parameter g and can reproduce either finite second moment distributions or infinite second moment distributions. In the proposed method, the real parameter q of the g-Gaussian distribution is encoded in the chromosome of an individual and is allowed to evolve. An evolutionary programming algorithm with the proposed idea is presented. Experiments were carried out to study the performance of the proposed algorithm.
Addressing dynamic optimization problems has attracted a growing interest from the evolutionary a... more Addressing dynamic optimization problems has attracted a growing interest from the evolutionary algorithm community in recent years due to its importance in the applications of evolutionary algorithms in real world problems. In order to study evolutionary algorithms in dynamic environments, one important work is to develop benchmark dynamic environments. This paper proposes two continuous dynamic problem generators. Both generators use linear transformation to move individuals, which preserves the distance among individuals. In the first generator, the linear transformation of individuals is equivalent to change the direction of some axes of the search space while in the second one it is obtained by successive rotations in different planes. Preliminary experiments were carried out to study the performance of some standard genetic algorithms in continuous dynamic environments created by the proposed generators.
This paper proposes a genetic algorithm (GA) with random immigrants for dynamic optimization prob... more This paper proposes a genetic algorithm (GA) with random immigrants for dynamic optimization problems where the worst individual and its neighbours are replaced every generation. In this GA, the individuals interact with each other and, when their fitness is close, as in the case where the diversity level is low, one single replacement can affect a large number of individuals. This simple approach can take the system to a kind of self-organization behavior, known as self-organized criticality (SOC), which is useful to maintain the diversity of the population in dynamic environments and hence allows the GA to escape from local optima when the problem changes. The experimental results show that the proposed GA presents the phenomenon of SOC.
The problem of fault tolerance in cooperative manipulators rigidly connected to a solid object is... more The problem of fault tolerance in cooperative manipulators rigidly connected to a solid object is addressed in this paper. Four faults are considered: free-swinging joint faults, locked joint faults, incorrect measured joint position, and incorrect measured joint velocity. The faults are first detected by a, fault detection and isolation system. Free-swinging and locked joint faults are isolated using artificial neural networks. The other faults are isolated based on the kinematic constraints imposed on the cooperative system. After the isolation of the faults, the control system is reconfigured. Control laws for the system with passive or locked joints are developed. Results of the fault tolerance system applied in simulations and in a real cooperative system are presented.
Faults in robotic manipulators can cause economic losses and serious damages. In the paper, two a... more Faults in robotic manipulators can cause economic losses and serious damages. In the paper, two artificial neural networks are employed to provide FDI to robotic manipulators. The first is a multilayer perceptron trained with backpropagation utilized to reproduce the dynamic of the manipulator and, so, generate the residual vector. The second is a radial basis function network employed to classify the residual vector and, thus, generate the fault isolation. As the system model is not employed, false alarms due to modeling errors are avoided. Two different algorithms are employed to train the last network. The first employs ridge regression (a regularization type) and the second uses forward selection (an algorithm for subset selection). Simulations in a two link manipulator evince that the FDI system can detect and isolate correctly faults that occur in nontrained trajectories
The problem of the control of cooperative manipulators with passive joints and are rigidly connec... more The problem of the control of cooperative manipulators with passive joints and are rigidly connected to a solid object is addressed in this paper. Passive joints can appear due to free-swinging joint failures or can be an intrinsic characteristic of the robots. A hybrid control of motion and squeeze force is proposed. For this purpose, a Jacobian matrix relating velocities in the actuated joints and load velocity is obtained based on the kinematic constraints of the cooperative system. Results of the control system applied in simulations and in real robots are presented.
Uploads
Papers by Renato Tinos