1. The complete title of one (or more) paper(s) published in the open literature describing the work that the author claims describes a human-competitive result. Automatic component-wise design of multiobjective evolutionary algorithms. 2. The name, complete physical mailing address, e-mail address, and phone number of EACH author of EACH paper: Leonardo C. T. Bezerra
IRIDIA, Université Libre de Bruxelles (ULB),
Av. F. Roosevelt 50, CP 194/6, 1050 Brussels, Belgium.
leonardo@iridia.ulb.ac.be
+32 (0) 2650 2729 Manuel López-Ibáñez
Alliance Manchester Business School
University of Manchester
Booth Street East, Manchester M13 9SS, UK
manuel.lopez-ibanez@manchester.ac.uk
+44 (0) 16130 68996 Thomas Stützle
IRIDIA, Université Libre de Bruxelles (ULB),
Av. F. Roosevelt 50, CP 194/6, 1050 Brussels, Belgium.
stuetzle@ulb.ac.be
+32 (0) 2650 3167 3. The name of the corresponding author (i.e., the author to whom notices will be sent concerning the competition) Leonardo Bezerra (leonardo@iridia.ulb.ac.be) 4. The abstract of the paper(s) Multiobjective evolutionary algorithms are typically proposed, studied and applied as monolithic blocks with a few numerical parameters that need to be set. Few works have studied how the algorithmic components of these evolutionary algorithms can be classified and combined to produce new algorithmic designs. The motivation for studies of this latter type stem from the development of flexible software frameworks and the usage of automatic algorithm configuration methods to find novel algorithm designs. In this paper, we propose a multiobjective evolutionary algorithm template and a new conceptual view of its components that surpasses existing frameworks in both the number of algorithms that can be instantiated from the template and the flexibility to produce novel algorithmic designs. We empirically demonstrate the flexibility of our proposed framework by automatically designing multiobjective evolutionary algorithms for continuous and combinatorial optimization problems. The automatically designed algorithms are often able to outperform six traditional multiobjective evolutionary algorithms from the literature, even after tuning their numerical parameters. 5. A list containing one or more of the eight letters (A, B, C, D, E, F, G, or H) that correspond to the criteria (see above) that the author claims that the work satisfies B The result is equal to or better than a result that was accepted as a new scientific result at the time when it was published in a peer-reviewed scientific journal. D The result is publishable in its own right as a new scientific result independent of the fact that the result was mechanically created. E The result is equal to or better than the most recent human-created solution to a long-standing problem for which there has been a succession of increasingly better human-created solutions. 6. A statement stating why the result satisfies the criteria that the
contestant claims. Multiobjective evolutionary algorithms have become the method of choice for tackling problems that present multiple (and often) conflicting objectives. In particular, the number of different MOEA proposals over the past decades is considerable, and the most significant advancements in the general multiobjective metaheuristics literature accredited to MOEA researchers is substantial. However, this vast MOEA literature has made the applicability of MOEAs to novel domains in an effective way a challenging task. More precisely, a practitioner wanting to adapt a MOEA to his target application problem is faced with a number of alternative MOEA proposals, but little information about how these algorithms perform on applications similar to his/her own. In this work, we proposed an effective, automated way to overcome this challenge. Specifically, we followed similar research on template-based automatic algorithm design and proposed a flexible MOEA template that can be used to automatically design effective MOEAs for given application domains. We empirically demonstrated the effectiveness of our methodology by automatically designing effective MOEAs for different application problems, ranging from continuous to combinatorial optimization. In fact, the automatically designed MOEAs consistently outperformed the manually designed ones from which their components had been selected. Therefore, our automatic design approach satisfies the following criteria: (B) Our automatically designed MOEAs are consistently better than the human-designed MOEAs, which had been accepted as novel scientific results at the time they were published in peer-reviewed scientific journals. (D) Our result was published in IEEE Transactions on Evolutionary Computation in its own right as a new scientific result independent of the fact that the result was mechanically created. Specifically, this high-impact journal is known for its rigorous reviewing process, and only accepted our work after assuring the effectiveness of the proposed approach. (E) Our result consistently outperforms some of the best human-created MOEAs both on continuous and combinatorial optimization. In particular, multiobjective continuous optimization is the primary application target of MOEAs and has been extensively studied over the past three decades. 7.  A full citation of the paper Leonardo C. T. Bezerra, M. López-Ibáñez and T. Stützle. The automatic component-wise design of multiobjective evolutionary algorithms. IEEE Transactions on Evolutionary Computation, 20(3):403–417, 2016. doi:10.1109/TEVC.2015.2474158. 8. Any prize money, if any, is to be divided equally among the co-authors. 9. A statement stating why the judges should consider the entry as "best" in comparison to other entries that may also be "human-competitive." The research on automatic algorithm design is broad and presents a number of successful case studies. In this sense, our work is one more effort in the direction of changing the traditional algorithm engineering methodology that has been used for decades in all optimization fields. However, our work is unique in a number of ways, as follows. 1. Compared to other works on hyperheuristics, our work has a much more challenging and significant scope. While hyperheuristics focus on the automatic design/selection of heuristics, our work focuses on the most relevant metaheuristic applied to multiobjective optimization. In particular, the structural complexity of MOEAs cannot be compared with that of devising heuristics. 2. Compared to other works on automatic design of multiobjective metaheuristics, our work is both more complex and pushes this research field further than any other. First, it is more complex as the literature on MOEAs far exceeds the literature on any other multiobjective metaheuristics. Second, it pushes this research field further since we address problems with any number of objectives, greatly enhancing the scope of our research. 3. Compared to other works on automatic algorithm design in general, our work has empirically demonstrated more practical applicability. For instance, most works on automatic design have targeted decision problems that share some structural similarities, such as satisfiability and qualified boolean formulas. Others have targeted combinatorial optimization, but have considered a single application problem. Our work is the first to simultaneously consider both continuous and combinatorial optimization, two domains where MOEAs have been repeatedly applied over the past years. Given the above arguments, we believe that our work goes beyond demonstrating that machine-created results can outperform human-created results, and has the concrete potential to become a groundbreaking work in the most relevant multiobjective metaheuristics research field.