Advances in compact Differential Evolution
نویسنده
چکیده
Compact Evolutionary Algorithms (cEAs) have been developed in order to address optimization problems characterized by limited memory resources. This situation is typical of robotics and control problems where a full power computing device may be unavailable due to cost and/or space limitations. These algorithms do not store and process an entire population and all its individuals therein but make use of a statistic representation of the population in order to perform the optimization process. In this way, a much smaller number of real numbers must be stored in the memory. The main drawback of these algorithms is that, since cEAs simulate the population by means of a statistical representation, they are subject to premature convergence, especially when the dimensionality grows. Nevertheless, among cEAs, compact Differential Evolution (cDE) has proven successful for a broad set of problems, but still it can be improved. In the first part of this talk I will describe several different novel cDE based algorithmic structures for solving continuous global optimization problems, proposed during my first research activities. These structures can be roughly divided in two categories: (1) structures which introduce additional components to standard cDE framework or (2) distributed structures which use cDE as a local search unit. In the first case the standard structure of cDE is modified, providing additional moves in order to further improve the search logic. In the latter case, several compact units evolve simultaneously and interact (with a global supervision or not) in order to solve the optimization problem, exploring the decision space from diverse perspectives. The structures introduced appear to be very promising as they allow the improvement of the results obtained by standard cDE. Numerical results show that the resulting algorithms proposed display a promising performance for a set of challenging test problems and they are competitive with the state-of-the-art algorithms. In the last part of this talk, I will introduce a possible real-world application of cDE based algorithms, specifically the minimization of the disturbance, due to motion, to the base of an anthropomorphic robotic arm installed on a spacecraft, by simulating the implementation of a compact algorithm within the robot micro-controller.
منابع مشابه
Fuzzy logic controlled differential evolution to solve economic load dispatch problems
In recent years, soft computing methods have generated a large research interest. The synthesis of the fuzzy logic and the evolutionary algorithms is one of these methods. A particular evolutionary algorithm (EA) is differential evolution (DE). As for any EA, DE algorithm also requires parameters tuning to achieve desirable performance. In this paper tuning the perturbation factor vector of DE ...
متن کاملFuzzy logic controlled differential evolution to solve economic load dispatch problems
In recent years, soft computing methods have generated a large research interest. The synthesis of the fuzzy logic and the evolutionary algorithms is one of these methods. A particular evolutionary algorithm (EA) is differential evolution (DE). As for any EA, DE algorithm also requires parameters tuning to achieve desirable performance. In this paper tuning the perturbation factor vector of DE ...
متن کاملIncreasing the Accuracy of Recommender Systems Using the Combination of K-Means and Differential Evolution Algorithms
Recommender systems are the systems that try to make recommendations to each user based on performance, personal tastes, user behaviors, and the context that match their personal preferences and help them in the decision-making process. One of the most important subjects regarding these systems is to increase the system accuracy which means how much the recommendations are close to the user int...
متن کاملA Compact Scheme for a Partial Integro-Differential Equation with Weakly Singular Kernel
Compact finite difference scheme is applied for a partial integro-differential equation with a weakly singular kernel. The product trapezoidal method is applied for discretization of the integral term. The order of accuracy in space and time is , where . Stability and convergence in norm are discussed through energy method. Numerical examples are provided to confirm the theoretical prediction ...
متن کاملDeveloping Adaptive Differential Evolution as a New Evolutionary Algorithm, Application in Optimization of Chemical Processes
متن کامل
Control of nonlinear systems using a hybrid APSO-BFO algorithm: An optimum design of PID controller
This paper proposes a novel hybrid algorithm namely APSO-BFO which combines merits of Bacterial Foraging Optimization (BFO) algorithm and Adaptive Particle Swarm Optimization (APSO) algorithm to determine the optimal PID parameters for control of nonlinear systems. To balance between exploration and exploitation, the proposed hybrid algorithm accomplishes global search over the whole search spa...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2010