Abstract Multi-objective evolutionary algorithms (MOEAs) are population based global optimization algorithms and it is said that the performance of the MOEAs depends on the population size. Considering that the recent trends of computer development is in large-scale many-core architectures, and massive parallel computation is getting feasible in more companies and laboratories, the available population size is increasing and the efficiency of... Multi-objective evolutionary algorithms (MOEAs) are population based global optimization algorithms and it is said that the performance of the MOEAs depends on the population size. Considering that the recent trends of computer development is in large-scale many-core architectures, and massive parallel computation is getting feasible in more companies and laboratories, the available population size is increasing and the efficiency of MOEA with large population size should be enhanced. This study examines the effect of the population size on MOEAs’ performance on a real-world-derived benchmarking optimization problem, with large population size. In this paper, three mate selection schemes with different degree of elitist strategy are adapted to NSGA-II-M2M. The experimental results show that the elitist strategy can efficiently make use of the effect of the large population size, therefore can reduce the turn-around time. Read More Read Less

Abstract Although the motivation to study multiobjective optimization algorithms comes from practice, there are only a few challenging real-world problems freely available to the research community. Because of this, algorithm benchmarking is performed primarily on artificial test problems. The most popular artificial test problems have characteristics that are not well-represented in real-world problems. This and the predominant inadequate performance assessment methodology widen the gap between... Although the motivation to study multiobjective optimization algorithms comes from practice, there are only a few challenging real-world problems freely available to the research community. Because of this, algorithm benchmarking is performed primarily on artificial test problems. The most popular artificial test problems have characteristics that are not well-represented in real-world problems. This and the predominant inadequate performance assessment methodology widen the gap between theory and practice in the eld of multiobjective optimization. The paper suggests to instead compare the algorithms with the anytime performance benchmarking approach of COCO (the Comparing Continuous Optimizers platform) on more realistic artificial problem suites as well as suites with diverse real-world problems. By listing the benets of sharing the real-world problems with the community, the paper hopes to encourage domain experts to embrace this practice. Read More Read Less

Abstract Humans have concerned with semantics behind words, objects and facts for thousands of years. Yet, when a computer uses a model and learns to recognize it from a video, does it really understand what it is or does it only map a number of specially colored pixels and shapes to a term “object”? In addition, humans use semantics to improve performance; therefore,... Humans have concerned with semantics behind words, objects and facts for thousands of years. Yet, when a computer uses a model and learns to recognize it from a video, does it really understand what it is or does it only map a number of specially colored pixels and shapes to a term “object”? In addition, humans use semantics to improve performance; therefore, it seems reasonable to assume that computers would as well. This paper overviews latest state of the art papers that learn and use semantics in order to improve the results of established methods in different areas. Two branches of AI: natural language processing and computer vision seem to be especially active in this area. Read More Read Less