Intervals in evolutionary algorithms for global optimization
Conference
·
OSTI ID:57622
Optimization is of central concern to a number of disciplines. Interval Arithmetic methods for global optimization provide us with (guaranteed) verified results. These methods are mainly restricted to the classes of objective functions that are twice differentiable and use a simple strategy of eliminating a splitting larger regions of search space in the global optimization process. An efficient approach that combines the efficient strategy from Interval Global Optimization Methods and robustness of the Evolutionary Algorithms is proposed. In the proposed approach, search begins with randomly created interval vectors with interval widths equal to the whole domain. Before the beginning of the evolutionary process, fitness of these interval parameter vectors is defined by evaluating the objective function at the center of the initial interval vectors. In the subsequent evolutionary process the local optimization process returns an estimate of the bounds of the objective function over the interval vectors. Though these bounds may not be correct at the beginning due to large interval widths and complicated function properties, the process of reducing interval widths over time and a selection approach similar to simulated annealing helps in estimating reasonably correct bounds as the population evolves. The interval parameter vectors at these estimated bounds (local optima) are then subjected to crossover and mutation operators. This evolutionary process continues for predetermined number of generations in the search of the global optimum.
- Research Organization:
- Los Alamos National Lab., NM (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- W-7405-ENG-36
- OSTI ID:
- 57622
- Report Number(s):
- LA-UR--95-1196; CONF-9503135--1-Extd.Abst.; ON: DE95010866
- Country of Publication:
- United States
- Language:
- English
Similar Records
A new hybrid algorithm combining a new chaos optimization approach with gradient descent for high dimensional optimization problems
Identification of ARMAX model for short term load forecasting: An evolutionary programming approach
Identification of ARMAX model for short term load forecasting: An evolutionary programming approach
Journal Article
·
Sun Jul 15 00:00:00 EDT 2018
· Computational and Applied Mathematics
·
OSTI ID:22769308
Identification of ARMAX model for short term load forecasting: An evolutionary programming approach
Journal Article
·
Wed Jan 31 23:00:00 EST 1996
· IEEE Transactions on Power Systems
·
OSTI ID:244774
Identification of ARMAX model for short term load forecasting: An evolutionary programming approach
Conference
·
Sat Dec 30 23:00:00 EST 1995
·
OSTI ID:367393