Búsqueda avanzada


Área de conocimiento




129 resultados, página 9 de 10

Control de sistemas usando aprendizaje de máquina

Systems control using machine learning

Jesús Martín Miguel Martínez (2023, [Tesis de maestría])

El aprendizaje por refuerzo es un paradigma del aprendizaje de máquina con un amplio desarrollo y una creciente demanda en aplicaciones que involucran toma de decisiones y control. Es un paradigma que permite el diseño de controladores que no dependen directamente del modelo que describe la dinámica del sistema. Esto es importante ya que en aplicaciones reales es frecuente que no se disponga de dichos modelos de manera precisa. Esta tesis tiene como objetivo implementar un controlador óptimo en tiempo discreto libre de modelo. La metodología elegida se basa en algoritmos de aprendizaje por refuerzo, enfocados en sistemas con espacios de estado y acción continuos a través de modelos discretos. Se utiliza el concepto de función de valor (Q-función y función V ) y la ecuación de Bellman para resolver el problema del regulador cuadrático lineal para un sistema mecánico masa-resorte-amortiguador, en casos donde se tiene conocimiento parcial y desconocimiento total del modelo. Para ambos casos las funciones de valor son definidas explícitamente por la estructura de un aproximador paramétrico, donde el vector de pesos del aproximador es sintonizado a través de un proceso iterativo de estimación de parámetros. Cuando se tiene conocimiento parcial de la dinámica se usa el método de aprendizaje por diferencias temporales en un entrenamiento episódico, que utiliza el esquema de mínimos cuadrados con mínimos cuadrados recursivos en la sintonización del crítico y descenso del gradiente en la sintonización del actor, el mejor resultado para este esquema es usando el algoritmo de iteración de valor para la solución de la ecuación de Bellman, con un resultado significativo en términos de precisión en comparación a los valores óptimos (función DLQR). Cuando se tiene desconocimiento de la dinámica se usa el algoritmo Q-learning en entrenamiento continuo, con el esquema de mínimos cuadrados con mínimos cuadrados recursivos y el esquema de mínimos cuadrados con descenso del gradiente. Ambos esquemas usan el algoritmo de iteración de política para la solución de la ecuación de Bellman, y se obtienen resultados de aproximadamente 0.001 en la medición del error cuadrático medio. Se realiza una prueba de adaptabilidad considerando variaciones que puedan suceder en los parámetros de la planta, siendo el esquema de mínimos cuadrados con mínimos cuadrados recursivos el que tiene los mejores resultados, reduciendo significativamente ...

Reinforcement learning is a machine learning paradigm with extensive development and growing demand in decision-making and control applications. This technique allows the design of controllers that do not directly depend on the model describing the system dynamics. It is useful in real-world applications, where accurate models are often unavailable. The objective of this work is to implement a modelfree discrete-time optimal controller. Through discrete models, we implemented reinforcement learning algorithms focused on systems with continuous state and action spaces. The concepts of value-function, Q-function, V -function, and the Bellman equation are employed to solve the linear quadratic regulator problem for a mass-spring-damper system in a partially known and utterly unknown model. For both cases, the value functions are explicitly defined by a parametric approximator’s structure, where the weight vector is tuned through an iterative parameter estimation process. When partial knowledge of the dynamics is available, the temporal difference learning method is used under episodic training, utilizing the least squares with a recursive least squares scheme for tuning the critic and gradient descent for the actor´s tuning. The best result for this scheme is achieved using the value iteration algorithm for solving the Bellman equation, yielding significant improvements in approximating the optimal values (DLQR function). When the dynamics are entirely unknown, the Q-learning algorithm is employed in continuous training, employing the least squares with recursive least squares and the gradient descent schemes. Both schemes use the policy iteration algorithm to solve the Bellman equation, and the system’s response using the obtained values was compared to the one using the theoretical optimal values, yielding approximately zero mean squared error between them. An adaptability test is conducted considering variations that may occur in plant parameters, with the least squares with recursive least squares scheme yielding the best results, significantly reducing the number of iterations required for convergence to optimal values.

aprendizaje por refuerzo, control óptimo, control adaptativo, sistemas mecánicos, libre de modelo, dinámica totalmente desconocida, aproximación paramétrica, Q-learning, iteración de política reinforcement learning, optimal control, adaptive control, mechanical systems, modelfree, utterly unknown dynamics, parametric approximation, Q-learning, policy iteration INGENIERÍA Y TECNOLOGÍA CIENCIAS TECNOLÓGICAS TECNOLOGÍA DE LOS ORDENADORES INTELIGENCIA ARTIFICIAL INTELIGENCIA ARTIFICIAL

Influence of conservation agriculture-based production systems on bacterial diversity and soil quality in rice-wheat-greengram cropping system in eastern Indo-Gangetic Plains of India

Anup Das virender kumar Peter Craufurd Andrew Mcdonald Sonam Sherpa (2023, [Artículo])

Introduction: Conservation agriculture (CA) is gaining attention in the South Asia as an environmentally benign and sustainable food production system. The knowledge of the soil bacterial community composition along with other soil properties is essential for evaluating the CA-based management practices for achieving the soil environment sustainability and climate resilience in the rice-wheat-greengram system. The long-term effects of CA-based tillage-cum-crop establishment (TCE) methods on earthworm population, soil parameters as well as microbial diversity have not been well studied. Methods: Seven treatments (or scenarios) were laid down with the various tillage (wet, dry, or zero-tillage), establishment method (direct-or drill-seeding or transplantation) and residue management practices (mixed with the soil or kept on the soil surface). The soil samples were collected after 7 years of experimentation and analyzed for the soil quality and bacterial diversity to examine the effect of tillage-cum-crop establishment methods. Results and Discussion: Earthworm population (3.6 times), soil organic carbon (11.94%), macro (NPK) (14.50–23.57%) and micronutrients (Mn, and Cu) (13.25 and 29.57%) contents were appreciably higher under CA-based TCE methods than tillage-intensive farming practices. Significantly higher number of OTUs (1,192 ± 50) and Chao1 (1415.65 ± 14.34) values were observed in partial CA-based production system (p ≤ 0.05). Forty-two (42) bacterial phyla were identified across the scenarios, and Proteobacteria, Actinobacteria, and Firmicutes were the most dominant in all the scenarios. The CA-based scenarios harbor a high abundance of Proteobacteria (2–13%), whereas the conventional tillage-based scenarios were dominated by the bacterial phyla Acidobacteria and Chloroflexi and found statistically differed among the scenarios (p ≤ 0.05). Composition of the major phyla, i.e., Proteobacteria, Actinobacteria, and Firmicutes were associated differently with either CA or farmers-based tillage management practices. Overall, the present study indicates the importance of CA-based tillage-cum-crop establishment methods in shaping the bacterial diversity, earthworms population, soil organic carbon, and plant nutrient availability, which are crucial for sustainable agricultural production and resilience in agro-ecosystem.

Metagenomics Bacterial Diversity Rice-Wheat-Greengram CIENCIAS AGROPECUARIAS Y BIOTECNOLOGÍA CONSERVATION AGRICULTURE DNA SEQUENCES EARTHWORMS METAGENOMICS SOIL QUALITY AGROECOSYSTEMS

Impact of automation on enhancing energy quality in grid-connected photovoltaic systems

Virgilio Alfonso Murillo Rodríguez NOE VILLA VILLASEÑOR José Manuel Robles Solís OA Guirette-Barbosa (2023, [Artículo])

Rapid growth in the integration of new consumers into the electricity sector, particularly in the industrial sector, has necessitated better control of the electricity supply and of the users’ op-erating conditions to guarantee an adequate quality of service as well as the unregulated dis-turbances that have been generated in the electrical network that can cause significant failures, breakdowns and interruptions, causing considerable expenses and economic losses. This research examines the characteristics of electrical variations in equipment within a company in the industrial sector, analyzes the impact generated within the electrical system according to the need for operation in manufacturing systems, and proposes a new solution through automation of the regulation elements to maintain an optimal system quality and prevent damage and equipment failures while offering a cost-effective model. The proposed solution is evaluated through a reliable simulation in ETAP (Energy Systems Modeling, Analysis and Optimization) software, which emulates the interaction of control elements and simulates the design of electric flow equipment operation. The results demonstrate an improvement in system performance in the presence of disturbances when two automation schemes are applied as well as the exclusive operation of the capacitor bank, which improves the total system current fluctuations and improves the power factor from 85.83% to 93.42%. Such a scheme also improves the waveform in the main power system; another improvement result is when simultaneously operating the voltage and current filter together with the PV system, further improving the current fluctuations, improving the power factor from 85.83% to 94.81%, achieving better stability and improving the quality of the waveform in the main power grid.

This article belongs to Special Issue Advances and Optimization of Electric Energy System.

Power quality Capacitor bank Voltage and current filter Photovoltaic system INGENIERÍA Y TECNOLOGÍA CIENCIAS TECNOLÓGICAS OTRAS ESPECIALIDADES TECNOLÓGICAS OTRAS OTRAS

Value chain research and development: The quest for impact

Jason Donovan (2023, [Artículo])

Motivation: For decades, governments, donors, and practitioners have promoted market-based development approaches (MBDA), most recently in the form of value chain development (VCD), to spur economic growth and reduce poverty. Changes in approaches have been shaped by funders, practitioners and researchers in ways that are incompletely appreciated. Purpose: We address the following questions: (1) how have researchers and practitioners shaped discussions on MBDA?; and (2) how has research stimulated practice, and how has practice informed research? We hypothesize that stronger exchange between researchers and practitioners increases the relevance and impact of value chain research and development. Methods and approach: We adopt Downs' (1972) concept of issue-attention cycles, which posits that attention to a particular issue follows a pattern where, first, excitement builds over potential solutions; followed by disenchantment as the inherent complexity, trade-offs, and resources required to solve it become apparent; and consequently attention moves on to a new issue. We review the literature on MBDA to see how far this framing applies. Findings: We identify five cycles of approaches to market-based development over the last 40 or more years: (1) non-traditional agricultural exports; (2) small and medium enterprise development; (3) value chains with a globalization perspective; (4) value chains with an agri-business perspective; and (5) value chain development. The shaping and sequencing of these cycles reflect researchers' tendency to analyse and criticize MBDA, while providing limited guidance on workable improvements; practitioners' reluctance to engage in critical reflection on their programmes; and an institutional and funding environment that encourages new approaches. Policy implications: Future MBDA will benefit from stronger engagement between researchers, practitioners, and funders. Before shifting attention to new concepts and approaches, achievements and failures in previous cycles need to be scrutinized. Evidence-based practice should extend for the length of the issue-attention cycle; preferably it should arrest the cycling of attention. Funders can help by requiring grantees to critically reflect on past action, by providing “safe spaces” for sharing such reflections, and by engaging in joint learning with practitioners and researchers.

Agri-Food Value Chains Issue-Attention Cycles Market-Based Development Approaches CIENCIAS AGROPECUARIAS Y BIOTECNOLOGÍA VALUE CHAINS PRIVATE SECTOR RURAL DEVELOPMENT SMALLHOLDERS