Applying extended automatic differentiation technique to process system optimization problems

2001 ◽  
Author(s):  
Li Xiang ◽  
Zhong Weitao ◽  
Shao Zhijiang ◽  
Qian Jixin
Keyword(s):  
Automatic Differentiation ◽  
Optimization Problems ◽  
System Optimization ◽  
Process System ◽  
Differentiation Technique

scholarly journals Gumbel-softmax-based optimization: a simple general framework for optimization problems on graphs

2021 ◽  
Vol 8 (1) ◽  
Author(s):  
Yaoxin Li ◽  
Jing Liu ◽  
Guozheng Lin ◽  
Yueyuan Hou ◽  
Muyun Mou ◽  
...  
Keyword(s):  
Objective Function ◽  
Statistical Physics ◽  
Automatic Differentiation ◽  
Optimization Problems ◽  
Vertex Cover ◽  
Independent Set ◽  
Computational Social Science ◽  
Maximization Problem ◽  
Maximum Independent Set ◽  
Gradient Descent Algorithm

AbstractIn computer science, there exist a large number of optimization problems defined on graphs, that is to find a best node state configuration or a network structure, such that the designed objective function is optimized under some constraints. However, these problems are notorious for their hardness to solve, because most of them are NP-hard or NP-complete. Although traditional general methods such as simulated annealing (SA), genetic algorithms (GA), and so forth have been devised to these hard problems, their accuracy and time consumption are not satisfying in practice. In this work, we proposed a simple, fast, and general algorithm framework based on advanced automatic differentiation technique empowered by deep learning frameworks. By introducing Gumbel-softmax technique, we can optimize the objective function directly by gradient descent algorithm regardless of the discrete nature of variables. We also introduce evolution strategy to parallel version of our algorithm. We test our algorithm on four representative optimization problems on graph including modularity optimization from network science, Sherrington–Kirkpatrick (SK) model from statistical physics, maximum independent set (MIS) and minimum vertex cover (MVC) problem from combinatorial optimization on graph, and Influence Maximization problem from computational social science. High-quality solutions can be obtained with much less time-consuming compared to the traditional approaches.

Optimum Selection of Hydraulic Devices for Water Hammer Control in the Pipeline Systems Using Genetic Algorithm

2003 ◽  
Author(s):  
Bong Seong Jung ◽  
Bryan W. Karney
Keyword(s):  
Genetic Algorithm ◽  
Steady State ◽  
Optimization Problems ◽  
System Optimization ◽  
Water Hammer ◽  
Pipeline System ◽  
Optimal Size ◽  
Protection Strategy ◽  
Hydraulic Devices ◽  
Selection Of

Genetic algorithms have been used to solve many water distribution system optimization problems, but have generally been limited to steady state or quasi-steady state optimization. However, transient events within pipe system are inevitable and the effect of water hammer should not be overlooked. The purpose of this paper is to optimize the selection, sizing and placement of hydraulic devices in a pipeline system considering its transient response. A global optimal solution using genetic algorithm suggests optimal size, location and number of hydraulic devices to cope with water hammer. This study shows that the integration of a genetic algorithm code with a transient simulator can improve both the design and the response of a pipe network. This study also shows that the selection of optimum protection strategy is an integrated problem, involving consideration of loading condition, device and system characteristics, and protection strategy. Simpler transient control systems are often found to outperform more complex ones.

scholarly journals Intelligent Total Transportation Management System for Future Smart Cities

2020 ◽  
Vol 10 (24) ◽  
pp. 8933
Author(s):  
Dinh Dung Nguyen ◽  
József Rohács ◽  
Dániel Rohács ◽  
Anita Boros
Keyword(s):  
Public Transport ◽  
Management System ◽  
Optimization Problems ◽  
Smart Cities ◽  
System Optimization ◽  
Preliminary Evaluation ◽  
Total System ◽  
Sensors And Actuators ◽  
Total Energy Consumption ◽  
Transportation Management

Smart mobility and transportation, in general, are significant elements of smart cities, which account for more than 25% of the total energy consumption related to smart cities. Smart transportation has seven essential sections: leisure, private, public, business, freight, product distribution, and special transport. From the management point of view, transportation can be classified as passive or non-cooperating, semi-active or simple cooperating, active or cooperating, contract-based, and priority transportation. This approach can be applied to public transport and even to passengers of public transport. The transportation system can be widely observed, analyzed, and managed using an extensive distribution network of sensors and actuators integrated into an Internet of Things (IoT) system. The paper briefly discusses the benefits that the IoT can offer for smart city transportation management. It deals with the use of a hierarchical approach to total transportation management, namely, defines the concept, methodology, and required sub-model developments, which describes the total system optimization problems; gives the possible system and methodology of the total transportation management; and demonstrates the required sub-model developments by examples of car-following models, formation motion, obstacle avoidances, and the total management system implementation. It also introduces a preliminary evaluation of the proposed concept relative to the existing systems.

scholarly journals Three dimensions, two microscopes, one code: Automatic differentiation for x-ray nanotomography beyond the depth of focus limit

2020 ◽  
Vol 6 (13) ◽  
pp. eaay3700 ◽  
Author(s):  
Ming Du ◽  
Youssef S. G. Nashed ◽  
Saugat Kandel ◽  
Doğa Gürsoy ◽  
Chris Jacobsen
Keyword(s):  
Multiple Scattering ◽  
Automatic Differentiation ◽  
Optimization Problems ◽  
Three Dimensions ◽  
Reconstruction Method ◽  
X Rays ◽  
Reconstruction Algorithms ◽  
Depth Of Focus ◽  
X Ray ◽  
Diffraction Effects

Conventional tomographic reconstruction algorithms assume that one has obtained pure projection images, involving no within-specimen diffraction effects nor multiple scattering. Advances in x-ray nanotomography are leading toward the violation of these assumptions, by combining the high penetration power of x-rays, which enables thick specimens to be imaged, with improved spatial resolution that decreases the depth of focus of the imaging system. We describe a reconstruction method where multiple scattering and diffraction effects in thick samples are modeled by multislice propagation and the 3D object function is retrieved through iterative optimization. We show that the same proposed method works for both full-field microscopy and for coherent scanning techniques like ptychography. Our implementation uses the optimization toolbox and the automatic differentiation capability of the open-source deep learning package TensorFlow, demonstrating a straightforward way to solve optimization problems in computational imaging with flexibility and portability.

On a Successive Approximation Technique in Solving Some Control System Optimization Problems

1963 ◽  
Vol 85 (2) ◽  
pp. 177-180 ◽  
Author(s):  
Masanao Aoki
Keyword(s):  
Successive Approximation ◽  
Optimization Problems ◽  
Structural Information ◽  
System Optimization ◽  
Extremal Problems ◽  
Approximation Technique ◽  
Minimizing Sequences ◽  
Approximation Techniques ◽  
Control Optimization ◽  
Computational Procedures

It has been realized for some time that most realistic optimization problems defy analytical solutions in closed forms and that in most cases it is necessary to resort to judicious combinations of analytical and computational procedures to solve problems. For example, in many optimization problems, one is interested in obtaining structural information on optimal and “good” suboptimal policies. Very often, various analytical as well as computational approximation techniques need be employed to obtain clear understandings of structures of policy spaces. The paper discusses a successive approximation technique to construct minimizing sequences for functionals in extremal problems, and the techniques will be applied, to a class of control optimization problems given by: Minv  J(v)=Minv  ∫01g(u.v)dt, where du/dt = h(u, v), h(u, v) linear in u and v, and where u and v are, in general, elements of Banach spaces. In Section 2, the minimizing sequences are constructed by approximating g(u, v) by appropriate quadratic expressions with linear constraining differential equations. It is shown that under the stated conditions the functional values converge to the minimal value monotonically. In Section 3, an example is included to illustrate some of the techniques discussed in the paper.

Sign in / Sign up

Export Citation Format

Share Document