Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 14 of 14 for “"Gradient Estimation"”.
-
W-SPSA : an Efficient Stochastic Approximation Algorithm for the off-line calibration of Dynamic Traffic Assignment models
… approaches focus on the separate or sequential estimation of demand and supply in a DTA system, a recently proposed framework calibrates the demand and supply models simultaneously by formulating the off-line calibration as a constrained optimization problem. Simultaneous Perturbation Stochastic …
-
Attention-Enabled Reinforcement Learning for Control of Scalable Multi-Agent Systems
… Specifically, on-policy algorithms based on gradient estimation have achieved state-of-the-art performance on end-to-end control problems once thought beyond the scope of machine learning methods. In seeking to apply the benefits of MARL to practical control of physical autonomous systems, we …
-
Global Optimization Techniques Based on Swarm-intelligent and Gradient-free Algorithms
… its saving in the computational power with the gradient-free attribute, are two solutions that deserve attention. Many researchers have exploited their merits in widely challenging applications. However, there is a known fact that both of them suffer from a severe drawback, non- effectively …
-
Online calibration for simulation-based dynamic traffic assignment : towards large-scale and real-time performance
… and congested networks; (3) presenting a gradient estimation procedure called partitioned simultaneous perturbation (PSP) that utilizes an assumed sparse gradient structure to facilitate real-time performance. The results demonstrate that, first, the proposed approach to address …
-
Sensitivities in Option Pricing Models
… the model and the market observations. Efficient gradient based optimization requires accurate gradient estimation of the cost function. In this thesis we highlight the adjoint method for computing gradients of the cost function in the context of gradient based optimization and show its …
-
Adaptive sampling trust-region methods for derivative-based and derivative-free simulation optimization problems
… In the first study we assume that the function gradients are directly observable through the Monte Carlo simulation. We propose ASTRO, which is an adaptive sampling based trust-region optimization method where a stochastic local model is constructed, optimized, and updated iteratively. ASTRO is …
-
Foundations of Multiple-Time-Scale Stochastic Approximation for Fast and Resilient Distributed Optimization Algorithms
… key aspects. First, we study consensus-based gradient methods over networks with a clustered structure, characterized by dense intra-cluster and sparse inter-cluster communication. This setting naturally gives rise to textit{two-time-scale} dynamics. We develop a framework to analyze these …
-
Frequency domain iterative tuning for active noise and vibration control
… and Vibration Control (ANVC). This approach is a gradient based self-tuning method which completely relies on analysis of the frequency response of system dynamics and the spectrum of signals. The new method is based on a new gradient estimation theory in the frequency domain. In this theory the …
-
Dynamic Modeling and Model-Free Real-Time Optimization for Cold Climate Heat Pump Systems
… By utilizing periodic dither inputs for online gradient estimation, ESC bears significant robustness against process variation and external disturbance, which has proved to be more advantageous in handling the challenging applications like heating, ventilation and air conditioning (HVAC) …
-
Laminar and Transitional Flow disturbances in Diseased and Stented Arteries
… novel method to improve the accuracy of velocity gradient estimation in the presence of noisy flow fields such as in DPIV where noise is inherently present is introduced with the objective to improve accuracy in the estimation of WSS, which are of paramount hemodynamic importance. The broad impact …
-
Learning-NUM: Utility Maximization in Stochastic Queueing Networks
… on function values to construct estimates on the gradients of the utility functions. The gradient estimates are then plugged into a first-order primal-dual framework, which is further embedded in a parallel-instance paradigm to deal with the feedback delay. We show that the resulting policy …
-
Anthropomimetic Control Synthesis: Adaptive Vehicle Traction Control
… driver would in order to maximize traction. A gradient ascent approach is at the heart of the proposed anthropomimetic control algorithm, and a real-time implementation is described using linear operator techniques, even though the tire-ground interface is highly non-linear. Performance of the …
-
A NEW ZEROTH-ORDER ORACLE FOR DISTRIBUTED AND NON-STATIONARY LEARNING
… explicit math- ematical form and therefore its gradient cannot be obtained. This invalidates all gradient-based optimization approaches. On the other hand, ZO methods approxi- mate the gradient by using the objective function values. Many existing ZO methods adopt the two-point feedback scheme …