Applications of genetic algorithms to optimization problems in the solvent extraction process for spent nuclear fuel
- Univ. of Tokyo (Japan). Dept. of Quantum Engineering and Systems Science
Applications of genetic algorithms (GAs) to optimization problems in the solvent extraction process for spent nuclear fuel are described. Genetic algorithms have been considered a promising tool for use in solving optimization problems in complicated and nonlinear systems because they require no derivatives of the objective function. In addition, they have the ability to treat a set of many possible solutions and consider multiple objectives simultaneously, so they can calculate many pareto optimal points on the trade-off curve between the competing objectives in a single iteration, which leads to small computing time. Genetic algorithms were applied to two optimization problems. First, process variables in the partitioning process were optimized using a weighted objective function. It was observed that the average fitness of a generation increased steadily as the generation proceeded and satisfactory solutions were obtained in all cases, which means that GAs are an appropriate method to obtain such an optimization. Secondly, GAs were applied to a multiobjective optimization problem in the co-decontamination process, and the trade-off curve between the loss of uranium and the solvent flow rate was successfully obtained. For both optimization problems, CPU time with the present method was estimated to be several tens of times smaller than with the random search method.
- Sponsoring Organization:
- USDOE
- OSTI ID:
- 477289
- Journal Information:
- Nuclear Technology, Journal Name: Nuclear Technology Journal Issue: 1 Vol. 118; ISSN 0029-5450; ISSN NUTYBB
- Country of Publication:
- United States
- Language:
- English
Similar Records
Multiobjective pressurized water reactor reload core design by nondominated genetic algorithm search
Multiobjective Optimal Controlled Variable Selection for a Gas Turbine–Solid Oxide Fuel Cell System Using a Multiagent Optimization Platform