scholarly journals Memory-efficient Genetic Algorithm for Path Optimization in Embedded Systems

2013 ◽  
Vol 6 (0) ◽  
pp. 28-36 ◽  
Author(s):  
Umair F. Siddiqi ◽  
Yoichi Shiraishi ◽  
Sadiq M. Sait
Keyword(s):  
Genetic Algorithm ◽  
Embedded Systems ◽  
Path Optimization ◽  
Memory Efficient

scholarly journals Robust CNN Compression Framework for Security-Sensitive Embedded Systems

2021 ◽  
Vol 11 (3) ◽  
pp. 1093
Author(s):  
Jeonghyun Lee ◽  
Sangkyun Lee
Keyword(s):  
Embedded Systems ◽  
Optimization Problem ◽  
State Of The Art ◽  
Classification Problems ◽  
Proximal Gradient Method ◽  
Knowledge Distillation ◽  
New Type ◽  
Adversarial Examples ◽  
Adversarial Training ◽  
Memory Efficient

Convolutional neural networks (CNNs) have achieved tremendous success in solving complex classification problems. Motivated by this success, there have been proposed various compression methods for downsizing the CNNs to deploy them on resource-constrained embedded systems. However, a new type of vulnerability of compressed CNNs known as the adversarial examples has been discovered recently, which is critical for security-sensitive systems because the adversarial examples can cause malfunction of CNNs and can be crafted easily in many cases. In this paper, we proposed a compression framework to produce compressed CNNs robust against such adversarial examples. To achieve the goal, our framework uses both pruning and knowledge distillation with adversarial training. We formulate our framework as an optimization problem and provide a solution algorithm based on the proximal gradient method, which is more memory-efficient than the popular ADMM-based compression approaches. In experiments, we show that our framework can improve the trade-off between adversarial robustness and compression rate compared to the existing state-of-the-art adversarial pruning approach.

scholarly journals Genetic algorithm for task mapping in embedded systems on a hierarchical architecture based on wireless network on chip WiNoC

DYNA ◽  
2017 ◽  
Vol 84 (201) ◽  
pp. 202 ◽  
Author(s):  
Maribell Sacanamboy Franco ◽  
Freddy Bolaños-Martinez ◽  
Álvaro Bernal-Noreña ◽  
Rubén Nieto-Londoño
Keyword(s):  
Genetic Algorithm ◽  
Embedded Systems ◽  
Wireless Network ◽  
Network On Chip ◽  
Hierarchical Architecture ◽  
Task Mapping ◽  
On Chip ◽  
La Red

Los sistemas de red en chip (NoC) fueron desarrollados originalmente para proporcionar un alto rendimiento, mediante la disponibilidad de varias unidades de procesamiento, conectadas a través de una red cableada dentro del circuito integrado. Wireless NoC (WiNoC o WNoC) son una evolución natural de los sistemas NoC, que integran una comunicación jerárquica dentro del chip para mejorar la escalabilidad. El mapeo de tareas en los sistemas WNoC representa un proceso desafiante, que a menudo implica varios objetivos de optimización, como potencia, rendimiento, productividad, uso de recursos y métricas de red. Este artículo describe un algoritmo genético basado en un enfoque para encontrar soluciones óptimas de asignación de tareas en tiempo de diseño, para sistemas embebidos que trabajan sobre un WiNoC. Los objetivos de optimización fueron: Aceleración, Consumo de Energía y Ancho de Banda. La red de destino utilizada para la simulación puede ser vista como un WiNoC jerárquica de dos niveles. El primer nivel corresponde a un conjunto de subredes que están conectadas por cables y son de tipo malla. El segundo nivel corresponde a una topología en estrella de enlaces inalámbricos, que conectan las subredes de primer nivel. El algoritmo propuesto muestra un buen desempeño en relación con los objetivos de optimización y la WiNoC heterogéneo simulada.

Genetic Algorithm Based Prefix Partitioning for Memory-Efficient IPv4/6 Lookup

2019 ◽  
Vol 17 (11) ◽  
pp. 1823-1830
Author(s):  
Fidel Ulises Sanchez Jimenez ◽  
Miguel Angel Ruiz Sanchez ◽  
Cesar Jalpa Villanueva
Keyword(s):  
Genetic Algorithm ◽  
Memory Efficient

scholarly journals Comparison of Genetic Algorithm and Hill Climbing for Shortest Path Optimization Mapping

2018 ◽  
Vol 31 ◽  
pp. 11017
Author(s):  
Mona Fronita ◽  
Rahmat Gernowo ◽  
Vincencius Gunawan
Keyword(s):  
Genetic Algorithm ◽  
Genetic Algorithms ◽  
Shortest Path ◽  
Path Optimization ◽  
Hill Climbing ◽  
Small Cities ◽  
Mutation Probability ◽  
Optimal Distance ◽  
Early Departure ◽  
Google Map Api

Traveling Salesman Problem (TSP) is an optimization to find the shortest path to reach several destinations in one trip without passing through the same city and back again to the early departure city, the process is applied to the delivery systems. This comparison is done using two methods, namely optimization genetic algorithm and hill climbing. Hill Climbing works by directly selecting a new path that is exchanged with the neighbour’s to get the track distance smaller than the previous track, without testing. Genetic algorithms depend on the input parameters, they are the number of population, the probability of crossover, mutation probability and the number of generations. To simplify the process of determining the shortest path supported by the development of software that uses the google map API. Tests carried out as much as 20 times with the number of city 8, 16, 24 and 32 to see which method is optimal in terms of distance and time computation. Based on experiments conducted with a number of cities 3, 4, 5 and 6 producing the same value and optimal distance for the genetic algorithm and hill climbing, the value of this distance begins to differ with the number of city 7. The overall results shows that these tests, hill climbing are more optimal to number of small cities and the number of cities over 30 optimized using genetic algorithms.

Sign in / Sign up

Export Citation Format

Share Document