This paper considers the problem of choosing optimal set (subset) of the descriptive variables (regressors) from a fixed set of candidates. Forward Selection and Backward Elimination methods adding/removing a candidate in/from the current set of descriptive variables step-by-step. Each variable is tested to be included or excluded using a chosen model comparison criteria that improves the model the most, and this process repeated until none improves the model. The model selection criteria may be calculated directly or recursively. Algorithms for recursive computing of the residuals sum of squares (RSS) for the model selection criteria in the recursive least squares method are presented. This paper evaluates the computational costs of the recursive calculation of stepwise model selection criteria for all possible steps of selection.
The task of designing the control actions for a heavy water reactor under uncertainty changes its parameters considered in the key differential game. The possibility of representing nonlinear dynamics of the object in the form of a system with parameters depending on the state (State Dependent Coefficients) and quadratic functional qualities allow you to go from having to solve a scalar partial differential equation (the Hamilton-Jacobi-Bellman) to the Riccati equation with parameters depending on the state. Feasible solution obtained by applying the min-max method. The results of mathematical modeling system in the shutdown of a nuclear reactor.
The paper deals with the algorithm of the identification of discrete systems with variable delay, consisting of an ideal sampler, zero-order hold and the linear continuous part. The delay parameter (fractional part of time delay) is estimated through the inverse modified Z-transform. The estimation is based on the equality of the continuous-time part step response to zero at the time delay point. The time delay of the discrete system (integer component) is adjusted by means of the integer part of the estimate obtained.
One of the widespread approaches to the issues of control in the group robotics is application of the social behavior models in the groups of robots. In this paper the author proposes to use this approach to fulfill the tasks of foraging. As a role model a Formicidae ant is proposed. This task is considered as a combination of three stages: finding food, returning to the ant hill and repeating the way to the place where food was found. It was proven that in order to come back home and repeatedly walk the way the Formicidae ants were navigated predominantly by the visual means using vector navigation (path integration) and landmark-guidance mechanisms. The basis of the proposed method is formed by the principles of memorizing the way by the visual landmarks and fuzzy control. The model of describing the way is introduced to the robot, which can define colors of the landmarks and approximately sense the direction to the landmark in respect to itself. A pattern for formation of a succinct way description was created, with the help of which the scout robot memorizes the way to the “food”. Certain regulations were developed, which let the follower robot transfer from the description of the route to the actions of its reproduction and in many ways copy an ant’s behavior. The actions are presented as elementary behavioral procedures, and each behavioral procedure is realized as a finite state automata. The results of the simulation modeling, which was conducted with the help of the framework of ROS based modeling system, are presented. Experiments were conducted in polygons with barriers and without them, with regular and irregular placing of the landmarks. As a quality criterion for the proposed method the author offers to consider a successful passing of the route by the follower robot, and this indicator in different series of experiments varies from 92 up to 98 %. The proposed method does not require robot’s great computation capacity and advanced sensory abilities. The developed method can also be applied to the tasks of reconnaissance and patrolling.
The necessity of using five ontologies such as the corporate culture ontology, the decision making ontology, the activity ontology, the facts ontology and the standard-legal ontology it is proved for constructing the intersubjective theories.
The computationally efficient method of fitness function evaluation (criterion for chromosomes selection) in genetic algorithms (GA) is discussed in this paper. This method may be used if a single gene modifies chromosome.
Steiner's problem in graphs is solved for the computing optimization. Population is represented as a weighted graph. Vertices of that graph represent chromosomes, edges represent the computational cost of selection criteria recurrent calculation. The GA application for identification of regression models assumes (a) gene is a regressor;
(b) chromosome is the set of regressors in single regression model (subset of all candidates);
(c) population — set of regression models (subset of all possible models); (d) selection criteria — residual sum of squares (RSS); (e) the chromosome modification by modification of one gene corresponds to the forward selection and backward elimination methods of variables (regressors) selection.
This paper considers a scheme of linearizing compensator for parallel Hammerstein model. The compensator exactly linearizes the system
via the internal feedback without truncation of Volterra series. Using the compensator with online identification determines adaptive linearization
This paper discusses local communication issues in a group of homogeneous robots for the purpose of decentralizing group management. A short review is presented of existing research in this area, which is mainly devoted to solutions for individual problems in the field. The possibility is considered to program the messages robots exchange within a group as fuzzy (pseudo-analog). There are comparisons with the natural world, where the social behavior of animals is negotiated with non-distinct messages in a continuous pattern. Issues regarding the physical aspects of organizing communication channels are considered. Robots that are used in group robotics have limited sensor and computing functions, but they should nonetheless be able to orient themselves relevant to one another to coordinate their common actions. Accordingly, the idea is proposed to emulate signal transmissions using a discrete IR-channel. The paper defends the grounds for interpreting received messages based on their sequence and the reactions they produce. The results of computer experiments that model the problem of individualized minds in robots are presented. The results of the computer experiments show that the use of fuzzy messages make robot behavior more variable, and allows the group to function more stably while consuming less energy for movement. These results prove that the proposed method is indeed viable, and also that message comprehension and the reliability of communication channels increases when fuzzy (pseudo-analog) messages are used.
In this paper we consider application of ant colony optimization techniques for capacitated vehicle routing problem. Modified ant colony optimization algorithm is proposed, computational results are reported.
The method of forming optimization algorithms for non-stationary control systems is developed in the article, based on the application of the Hamilton-Jacobi equation and the Pontryagin minimum principle. In this article, the original nonlinear differential equation that describes the original control system is transformed into a system with a linear structure, but with State Dependent Coefficient (SDC) parameters. The use of the quadratic quality criterion in problems with unlimited time of the transient process makes it possible, in the synthesis of control for the transformed system, to move from the need to search for the solution of a scalar partial differential equation (the Hamilton-Jacobi-Bellman equation) to a Riccati-type equation with state-dependent parameters. However, solving the resulting equation in the rate of the object's operation is no less difficult. For its solution, an algorithmic method for the synthesis of controls is proposed. The behavior of the Hamiltonian under optimal control changes during the transient process along a well-defined trajectory. This property of the Hamiltonian was used as the basis for the design of algorithms for optimizing the control system. When the formulated conditions are met, a "transfer" of the quality functional from peripheral values to its minimum value is guaranteed asymptotically. The effectiveness of the developed algorithms is demonstrated by the example of the synthesis of control controlling the supply of antiretroviral drugs HAART to the human body in the presence of HIV. The simulation was carried out in the MATLAB package.