Eugenius Kaszkurewicz - Academia.edu (original) (raw)
Papers by Eugenius Kaszkurewicz
1993 American Control Conference, 1993
This paper describes an implementation of the Diagnostic Model Processor methodology for fault an... more This paper describes an implementation of the Diagnostic Model Processor methodology for fault analysis of a gas-oil separation stage derived from the design of a real oil production plant in an offshore platform. The theory of the method is reviewed and two examples of the diagnostic system's performance are given. One of the examples exposes some problems of the method when applied to a process with few sensors. Some suggestions for improvements are considered, specially with respect to the multiple fault identification problem. The prototype system studied was supplied with data from a simulator and was operated off-line.
Complexity, 2020
This paper extends the Deal-Vidal-Wolfe and Lanchester models of duopoly dynamics, which involve ... more This paper extends the Deal-Vidal-Wolfe and Lanchester models of duopoly dynamics, which involve two populations, by explicitly introducing a third population of undecided users. An analysis of these extended models establishes conditions for the existence of equilibria, as well as their stability properties under different classes of advertising policies. This analysis also leads to the surprising result that the extended Vidale–Wolfe and Lanchester models, despite having different dynamics, under the general class of decentralized affine feedback advertising policies have equilibria in identical locations, with the same stability properties.
IFAC Proceedings Volumes, 1993
This paper describes the prototype of a knowledge-based system designed for fault detection and d... more This paper describes the prototype of a knowledge-based system designed for fault detection and diagnosis in oil production plants in offshore platforms. This prototype uses a Goal Tree-Success Tree (GTST) model to represent a gas-oil separation process, derived from the design of an actual platform. Process data for the development and tests of the prototype are generated by a dynamical simulator that reproduces the plant behavior under normal and faulty conditions. The results obtained are presented and perspectives for the implementation of the system in a rea.! ca.", are discussed.
IFAC Proceedings Volumes, 2002
Threshold policies are defined and analysed for different t i p of (in(species and predator-prey ... more Threshold policies are defined and analysed for different t i p of (in(species and predator-prey type models. It is shown that such policie' l i t designed by suitable choice of so called virtual equilibrium points. The ~11np1(st threshold policies are discontinuous, which leads to some drawbatks It 1-, 1 1 9~ shown how to design continuous threshold policies that retain niost CJf rhc advantages of their discontinuous versions but do not have the majoi di au back of chatter in the control. Threshold policies are also seen to be robust t o unccwaintv of model parameters, and initial conditions, as well as t o delays bPtaeeii stock assessment and policy enactment. The models which are controlled are the1 Noy-Meir herbivore-vegetation model, subject to linear consumption curves, the. two species Rosenzweig-MacArthur model, and a three-dimensional chemostat model.
Sba: Controle & Automação Sociedade Brasileira de Automatica, 2005
A large class of predator-prey models can be written as a nonlinear dynamical system in one or tw... more A large class of predator-prey models can be written as a nonlinear dynamical system in one or two variables (species). In many contexts, it is necessary to introduce a control into these dynamics. In this paper we focus on models of two species, and assume, as is common in mathematical ecology, that the control corresponds to a proportional removal of the predator population. Six controller design techniques are applied to the Lotka-Volterra model, which is thus used as a benchmark to evaluate and compare these techniques in an ecological context.
2014 IEEE Congress on Evolutionary Computation (CEC), 2014
This paper introduces an ACO model and associated algorithm, called Coupled EigenAnt, for the pro... more This paper introduces an ACO model and associated algorithm, called Coupled EigenAnt, for the problem of finding the shortest of N paths between a source and a destination node. It is based on the recently introduced EigenAnt algorithm, the novelty being that it allows probabilistic path choice on both the forward and return journeys, as well as the fact that it introduces decay of pheromone deposition following a geometric progression. Equilibrium points of the model are calculated and the local stability of the two path synchronous version analyzed. Simulations illustrate the main features of the algorithm.
Proceedings of the 41st IEEE Conference on Decision and Control, 2002.
Abstract Proposes the control of predator-prey type models through the use of on-off control, als... more Abstract Proposes the control of predator-prey type models through the use of on-off control, also known as a threshold policy in the biological context. Both discontinuous control, similar to variable structure control, and a continuous version of this control are used. Robustness of the proposed controls is discussed.
Theoretical Population Biology, 2005
The stability of predator-prey models, in the context of exploitation of renewable resources, sub... more The stability of predator-prey models, in the context of exploitation of renewable resources, subject to threshold policies (TP) is studied in this paper using the idea of backstepping and Control Liapunov Functions (CLF) well known in control theory, as well as the concept of virtual equilibria. TPs are defined and analysed for different types of one and two species predator-prey models. The models studied are the single species Noy-Meir herbivore-vegetation model, in a grazing management context, as well as the Rosenzweig-MacArthur two species predator-prey model, in a fishery management context. TPs are shown to be versatile and useful in managing renewable resources, being simple to design and implement, and also yielding advantages in situations of overexploitation.
Nonlinear Analysis: Real World Applications, 2006
In this paper, the control of a ratio-dependent predator-prey system is considered. This system, ... more In this paper, the control of a ratio-dependent predator-prey system is considered. This system, in its equivalent nondimensional form, is subjected to an on-off control, also known as a threshold policy. The control, applied only to the predator, is of the "constant effort removal" type. It is shown that such a control can be designed by a suitable choice of so called virtual equilibrium points. The analysis of the global stability of a specific nondimensional model is carried out. It shows that, in the nondimensional model, the on-off control changes the original equilibrium points on the axes to a globally stable positive equilibrium point, attained through a sliding mode, thus avoiding extinction, and achieving coexistence of both species.
IEEE Transactions on Parallel and Distributed Systems, 1996
IEEE Transactions on Automatic Control, 2007
Abstract In this paper, it is shown how standard iterative methods for solving linear and nonline... more Abstract In this paper, it is shown how standard iterative methods for solving linear and nonlinear equations can be designed from the point of view of control. Appropriate choices of control Lyapunov functions (CLFs) lead to both continuous and discrete-time versions of the Newton-Raphson and conjugate gradient algorithms as well as new variants.
Expert Systems with Applications, 1997
This paper describes an expert system for process supervision and diagnosis of an offshore oil pr... more This paper describes an expert system for process supervision and diagnosis of an offshore oil product plant. The expert system contains main modules with detection, diagnostic and advisory functions. The implementation issues of the expert system using the Gensym G2 software environment are discussed. The system was developed on board the Petrobras XXIV platform in the Campos Basin, Rio de Janeiro, Brazil.
Ecological Modelling, 2000
This paper analyses the dynamics of two predator-prey models, namely, Lotka-Volterra and Leslie-G... more This paper analyses the dynamics of two predator-prey models, namely, Lotka-Volterra and Leslie-Gower, when submitted to a specific harvesting schedule known as the weighted escapement policy. The main goal is to show that global stability of a desired equilibrium population can be achieved by the application of simple switching controls, chosen in accordance with variable structure system theory. A qualitative analysis shows that the global stability depends on the existence of virtual equilibrium points which in turn is ensured by a suitable choice of the control parameters.
Ecological Engineering, 2006
In the grazing management context, an on-off policy as well as a hysteresis on-off policy can be ... more In the grazing management context, an on-off policy as well as a hysteresis on-off policy can be used to avoid the collapse in the vegetation-herbivore dynamical system model. We consider a dynamical system sensitive to soil degradation. The on-off policy changes the dynamics of the vegetation-herbivore system in such a way that a positive stable equilibrium point is achieved, and the hysteresis on-off policy changes the dynamics in such a way that a limit cycle (bounded oscillation) can be achieved, and the multiple stable states that arise in this type of system are avoided. A new model is introduced in which changes in the dynamics that occur on account of switching are explicitly introduced. Policies can be designed by a suitable choice of so called virtual equilibrium points in a simple and intuitive manner.
Resumo—Este artigo apresenta diversos algoritmos contınuos caracterizados por equaçoes diferencia... more Resumo—Este artigo apresenta diversos algoritmos contınuos caracterizados por equaçoes diferenciais ordinárias (EDO) de segunda ordem, os quais podem ser interpretados como versoes contınuas do conhecido algoritmo discreto de gradiente conjugado (CG). O objetivo destes algoritmos é, dada uma funçao objetivo escalar nao convexa, gerar trajetórias contınuas que, partindo de um ponto inicial arbitrário, sejam capazes de ultrapassar mınimos locais para se estacionarem no mınimo global desconhecido da funçao objetivo. ...
Resumo—O problema de restaurar imagens degradadas por distorções lineares e invariantes no espaço... more Resumo—O problema de restaurar imagens degradadas por distorções lineares e invariantes no espaço e afetadas por ruído é resolvido através de uma regularização em norma L1, que equivale a determinar uma solução em norma L1 de um sistema sobredeterminado de equações lineares, que resulta da norma L1 do erro somado a um termo de regularização, também em norma L1. Este sistema é resolvido através de um sistema gradiente com o segundo membro descontínuo, que pode ser interpretado como ...
Resumo Este trabalho apresenta uma solução para o problema de carregamento de container através d... more Resumo Este trabalho apresenta uma solução para o problema de carregamento de container através de algoritmo genético, utilizando algumas características reais e adaptadas do processo de carregamento e dos dados das cargas de uma empresa de produto eletrônico. O algoritmo desenvolvido propõe dois modelos de carregamento (preenchimento do container), unificando duas propostas feitas anteriormente na literatura. São realizados testes com os dois modelos de carregamento com diferentes ...
1993 American Control Conference, 1993
This paper describes an implementation of the Diagnostic Model Processor methodology for fault an... more This paper describes an implementation of the Diagnostic Model Processor methodology for fault analysis of a gas-oil separation stage derived from the design of a real oil production plant in an offshore platform. The theory of the method is reviewed and two examples of the diagnostic system's performance are given. One of the examples exposes some problems of the method when applied to a process with few sensors. Some suggestions for improvements are considered, specially with respect to the multiple fault identification problem. The prototype system studied was supplied with data from a simulator and was operated off-line.
Complexity, 2020
This paper extends the Deal-Vidal-Wolfe and Lanchester models of duopoly dynamics, which involve ... more This paper extends the Deal-Vidal-Wolfe and Lanchester models of duopoly dynamics, which involve two populations, by explicitly introducing a third population of undecided users. An analysis of these extended models establishes conditions for the existence of equilibria, as well as their stability properties under different classes of advertising policies. This analysis also leads to the surprising result that the extended Vidale–Wolfe and Lanchester models, despite having different dynamics, under the general class of decentralized affine feedback advertising policies have equilibria in identical locations, with the same stability properties.
IFAC Proceedings Volumes, 1993
This paper describes the prototype of a knowledge-based system designed for fault detection and d... more This paper describes the prototype of a knowledge-based system designed for fault detection and diagnosis in oil production plants in offshore platforms. This prototype uses a Goal Tree-Success Tree (GTST) model to represent a gas-oil separation process, derived from the design of an actual platform. Process data for the development and tests of the prototype are generated by a dynamical simulator that reproduces the plant behavior under normal and faulty conditions. The results obtained are presented and perspectives for the implementation of the system in a rea.! ca.", are discussed.
IFAC Proceedings Volumes, 2002
Threshold policies are defined and analysed for different t i p of (in(species and predator-prey ... more Threshold policies are defined and analysed for different t i p of (in(species and predator-prey type models. It is shown that such policie' l i t designed by suitable choice of so called virtual equilibrium points. The ~11np1(st threshold policies are discontinuous, which leads to some drawbatks It 1-, 1 1 9~ shown how to design continuous threshold policies that retain niost CJf rhc advantages of their discontinuous versions but do not have the majoi di au back of chatter in the control. Threshold policies are also seen to be robust t o unccwaintv of model parameters, and initial conditions, as well as t o delays bPtaeeii stock assessment and policy enactment. The models which are controlled are the1 Noy-Meir herbivore-vegetation model, subject to linear consumption curves, the. two species Rosenzweig-MacArthur model, and a three-dimensional chemostat model.
Sba: Controle & Automação Sociedade Brasileira de Automatica, 2005
A large class of predator-prey models can be written as a nonlinear dynamical system in one or tw... more A large class of predator-prey models can be written as a nonlinear dynamical system in one or two variables (species). In many contexts, it is necessary to introduce a control into these dynamics. In this paper we focus on models of two species, and assume, as is common in mathematical ecology, that the control corresponds to a proportional removal of the predator population. Six controller design techniques are applied to the Lotka-Volterra model, which is thus used as a benchmark to evaluate and compare these techniques in an ecological context.
2014 IEEE Congress on Evolutionary Computation (CEC), 2014
This paper introduces an ACO model and associated algorithm, called Coupled EigenAnt, for the pro... more This paper introduces an ACO model and associated algorithm, called Coupled EigenAnt, for the problem of finding the shortest of N paths between a source and a destination node. It is based on the recently introduced EigenAnt algorithm, the novelty being that it allows probabilistic path choice on both the forward and return journeys, as well as the fact that it introduces decay of pheromone deposition following a geometric progression. Equilibrium points of the model are calculated and the local stability of the two path synchronous version analyzed. Simulations illustrate the main features of the algorithm.
Proceedings of the 41st IEEE Conference on Decision and Control, 2002.
Abstract Proposes the control of predator-prey type models through the use of on-off control, als... more Abstract Proposes the control of predator-prey type models through the use of on-off control, also known as a threshold policy in the biological context. Both discontinuous control, similar to variable structure control, and a continuous version of this control are used. Robustness of the proposed controls is discussed.
Theoretical Population Biology, 2005
The stability of predator-prey models, in the context of exploitation of renewable resources, sub... more The stability of predator-prey models, in the context of exploitation of renewable resources, subject to threshold policies (TP) is studied in this paper using the idea of backstepping and Control Liapunov Functions (CLF) well known in control theory, as well as the concept of virtual equilibria. TPs are defined and analysed for different types of one and two species predator-prey models. The models studied are the single species Noy-Meir herbivore-vegetation model, in a grazing management context, as well as the Rosenzweig-MacArthur two species predator-prey model, in a fishery management context. TPs are shown to be versatile and useful in managing renewable resources, being simple to design and implement, and also yielding advantages in situations of overexploitation.
Nonlinear Analysis: Real World Applications, 2006
In this paper, the control of a ratio-dependent predator-prey system is considered. This system, ... more In this paper, the control of a ratio-dependent predator-prey system is considered. This system, in its equivalent nondimensional form, is subjected to an on-off control, also known as a threshold policy. The control, applied only to the predator, is of the "constant effort removal" type. It is shown that such a control can be designed by a suitable choice of so called virtual equilibrium points. The analysis of the global stability of a specific nondimensional model is carried out. It shows that, in the nondimensional model, the on-off control changes the original equilibrium points on the axes to a globally stable positive equilibrium point, attained through a sliding mode, thus avoiding extinction, and achieving coexistence of both species.
IEEE Transactions on Parallel and Distributed Systems, 1996
IEEE Transactions on Automatic Control, 2007
Abstract In this paper, it is shown how standard iterative methods for solving linear and nonline... more Abstract In this paper, it is shown how standard iterative methods for solving linear and nonlinear equations can be designed from the point of view of control. Appropriate choices of control Lyapunov functions (CLFs) lead to both continuous and discrete-time versions of the Newton-Raphson and conjugate gradient algorithms as well as new variants.
Expert Systems with Applications, 1997
This paper describes an expert system for process supervision and diagnosis of an offshore oil pr... more This paper describes an expert system for process supervision and diagnosis of an offshore oil product plant. The expert system contains main modules with detection, diagnostic and advisory functions. The implementation issues of the expert system using the Gensym G2 software environment are discussed. The system was developed on board the Petrobras XXIV platform in the Campos Basin, Rio de Janeiro, Brazil.
Ecological Modelling, 2000
This paper analyses the dynamics of two predator-prey models, namely, Lotka-Volterra and Leslie-G... more This paper analyses the dynamics of two predator-prey models, namely, Lotka-Volterra and Leslie-Gower, when submitted to a specific harvesting schedule known as the weighted escapement policy. The main goal is to show that global stability of a desired equilibrium population can be achieved by the application of simple switching controls, chosen in accordance with variable structure system theory. A qualitative analysis shows that the global stability depends on the existence of virtual equilibrium points which in turn is ensured by a suitable choice of the control parameters.
Ecological Engineering, 2006
In the grazing management context, an on-off policy as well as a hysteresis on-off policy can be ... more In the grazing management context, an on-off policy as well as a hysteresis on-off policy can be used to avoid the collapse in the vegetation-herbivore dynamical system model. We consider a dynamical system sensitive to soil degradation. The on-off policy changes the dynamics of the vegetation-herbivore system in such a way that a positive stable equilibrium point is achieved, and the hysteresis on-off policy changes the dynamics in such a way that a limit cycle (bounded oscillation) can be achieved, and the multiple stable states that arise in this type of system are avoided. A new model is introduced in which changes in the dynamics that occur on account of switching are explicitly introduced. Policies can be designed by a suitable choice of so called virtual equilibrium points in a simple and intuitive manner.
Resumo—Este artigo apresenta diversos algoritmos contınuos caracterizados por equaçoes diferencia... more Resumo—Este artigo apresenta diversos algoritmos contınuos caracterizados por equaçoes diferenciais ordinárias (EDO) de segunda ordem, os quais podem ser interpretados como versoes contınuas do conhecido algoritmo discreto de gradiente conjugado (CG). O objetivo destes algoritmos é, dada uma funçao objetivo escalar nao convexa, gerar trajetórias contınuas que, partindo de um ponto inicial arbitrário, sejam capazes de ultrapassar mınimos locais para se estacionarem no mınimo global desconhecido da funçao objetivo. ...
Resumo—O problema de restaurar imagens degradadas por distorções lineares e invariantes no espaço... more Resumo—O problema de restaurar imagens degradadas por distorções lineares e invariantes no espaço e afetadas por ruído é resolvido através de uma regularização em norma L1, que equivale a determinar uma solução em norma L1 de um sistema sobredeterminado de equações lineares, que resulta da norma L1 do erro somado a um termo de regularização, também em norma L1. Este sistema é resolvido através de um sistema gradiente com o segundo membro descontínuo, que pode ser interpretado como ...
Resumo Este trabalho apresenta uma solução para o problema de carregamento de container através d... more Resumo Este trabalho apresenta uma solução para o problema de carregamento de container através de algoritmo genético, utilizando algumas características reais e adaptadas do processo de carregamento e dos dados das cargas de uma empresa de produto eletrônico. O algoritmo desenvolvido propõe dois modelos de carregamento (preenchimento do container), unificando duas propostas feitas anteriormente na literatura. São realizados testes com os dois modelos de carregamento com diferentes ...