Books

1.

Yu, Jun Takagi Hideyuki, Explosion Operation of Fireworks Algorithm, IGI Global, 10.4018/978-1-7998-1659-1, Jun Yu, Hideyuki Takagi, "Explosion Operation of Fireworks Algorithm," Chapter 3 of Handbook of Research on Fireworks Algorithms and Swarm Intelligence, (ed.) Ying Tan, IGI Global (2020). DOI: 10.4018/978-1-7998-1659-1, 2020.05, [URL], This chapter briefly reviews the basic explosion mechanism used in the fireworks algorithm (FWA) and comprehensively investigates relevant research on explosion operations. Since the explosion mechanism is one of the most core operations directly affecting the performance of FWA, the authors focus on analyzing the FWA explosion operation and highlighting two novel explosion strategies: a multi-layer explosion strategy and a scouting explosion strategy. The multi-layer explosion strategy allows an individual firework to perform multiple explosions instead of the single explosion used in the original FWA, where each round of explosion can be regarded as a layer; the scouting explosion strategy controls an individual firework to generate spark individuals one by one instead of generating all spark individuals within the explosion amplitude at once. The authors then introduce several other effective strategies to further improve the performance of FWA by full using the information generated by the explosion operation. Finally, the authors list some open topics for discussion..

Reports

Papers

1.	Hideyuki Takagi and Miho Ohsaki, Interactive Evolutionary Computation-Based Hearing-Aid Fitting, IEEE Transaction on Evolutionary Computation, vol.11, no.3, pp.414-427 (2007), 2007.06.

Presentations

1.	YU, Jun, TAKAGI, Hideyuki, Accelerating Fireworks Algorithm with Dynamic Population Size Strategy, Joint 11th International Conference on Soft Computing and Intelligent Systems and 21st International Symposium on Advanced Intelligent Systems (SCIS&ISIS2020), 2020.12, A dynamic population size strategy is proposed for the fireworks algorithm (FWA) to adjust the population size based to the search results of the current generation. When the currently found optimal individual is updated, a linear decreasing method is activated to maintain an efficient exploitation speed. The population size is reduced by 1 until the minimum pre-set population size is reached, then the population size remains unchanged. Otherwise, we randomly generate a larger population size than the initial population and expand the explosion amplitudes of all firework individuals artificially, which the expectation that we can escape current local minima. To analyze the effectiveness of the proposed strategy, we combined it with the enhanced FWA (EFWA) together, and run the EFWA and (the EFWA + our proposed strategy) on 28 CEC 2013 benchmark functions in three different dimensions. Each function is run 30 trial times independently, and the Wilcoxon signed-rank test is applied to check significant differences. The statistical results showed that the proposed dynamic population size strategy can not only achieve a faster convergence speed for the FWA but also can jump out of trapped local minima more easily to maintain a higher performance, especially for high-dimensional problems..
2.	INOUE, Makoto, MATSUMOTO, Hibiki, TAKAGI, Hideyuki, Acceptability of a Decision Maker to Handle Multi-objective Optimization on Design Space, Joint 11th International Conference on Soft Computing and Intelligent Systems and 21st International Symposium on Advanced Intelligent Systems (SCIS&ISIS2020), 2020.12, We introduce the acceptability of a decision maker to handle evolutionary multi-objective optimization (EMO) on design space, while most of EMO research tries to find many solutions on an objective space and passes them to a decision maker. Unlike this conventional EMO approaches, our approach decides maker's model with the concept of acceptability and introduces it in EMO search. Especially, this approach works well when qualitative factors, such as the decision maker's experience and knowledge on a task, are a part of evaluations. Acceptability functions for each of objectives are aggregated firstly, and the aggregated acceptability forms contours on an objective space and is mapped on a design space. The acceptability contours on a design space can narrow down the area of solutions. We could find better solutions in our experiments than the conventional approach of searching solutions on an objective space..
3.	YU, Jun, TAKAGI, Hideyuki, Multi-species Generation Strategy Based Vegetation Evolution, 2020 IEEE Congress on Evolutionary Computation (CEC2020), 2020.07, We propose a multi-species generation strategy to increase the diversity of seed individuals produced in the maturity operation of vegetation evolution (VEGE). Since the breeding patterns of real plants can be roughly divided into sexual reproduction and asexual one, the proposed strategy additionally introduces two different methods to simulate these two patterns. As our preliminary attempt of the simulation, a mature individual is splattered randomly in the neighbor local area of its parent individual with Gaussian distribution probability to simulate asexual reproduction, while a mature individual is generated by crossing randomly selected two different parent individuals to simulate sexual reproduction. Our proposed strategy consists of these two new reproduction methods and that of our original VEGE, and each mature individual in every generation randomly selects one of these three methods to generate seed individuals, which is analogous to different plant species using different mechanisms to breed. To evaluate the performance of our proposed strategy, we compare VEGE and (VEGE + the proposed generation strategy) on 28 benchmark functions of three different dimensions from the CEC 2013 test suit with 30 independent trial runs. The experimental results have confirmed that the proposed strategy can increase the diversity of seed individuals, accelerate the convergence of VEGE significantly, and become effective according to the increase of dimensions..
4.	LI, Yuhao, YU, Jun, TAKAGI, Hideyuki, Niche Method Complementing the Nearest-better Clustering, 2019 IEEE Symposium Series on Computational Intelligence (SSCI2019), 2019.12, We propose a two-stage niching algorithm that separates local optima areas in the first stage and finds the optimum point of each area using any optimization technique in the second stage. The proposed first stage has complementary characteristics to the shortcoming of Nearest-better Clustering (NBC). We introduce a weighted gradient and distance-based clustering method (WGraD) and two methods for determining its weights to find out niches and overcome NBC. The WGraD creates spanning trees by connecting each search point to other suitable one decided by weighted gradient information and weighted distance information among search points. Since weights influence its clustering result, we propose two weight determination methods 1 and 2. The weight determination method 1 We combine these methods into WGrad, i.e. WGraD1 and WGraD2, and compare the characteristics of NBC, WGraD1, and WGraD2 using differential evolution (DE) as a baseline search algorithm for obtaining the optimum of each niche after clustering local areas. We design a controlled experiment and run (NBC + DE), (WGraD1 + DE) and (WGraD2 + DE) on 8 benchmark functions from CEC 2015 test suite for single objective multiniche optimization. The experimental results confirmed that the proposed strategy can overcome the shortcoming of NBC and be a complementary niche method of NBC. .
5.	YU, Jun, TAKAGI, Hideyuki, Accelerating Vegetation Evolution with Mutation Strategy and Gbased Growth Strategy, 2019 IEEE Symposium Series on Computational Intelligence (SSCI2019), 2019.12, We propose two strategies, mutation strategy and Gbased growth strategy, to enhance the performance of standard vegetation evolution (VEGE) that simulates the growth and reproduction of vegetation repeatedly to find the global optimum. We introduce two different mutation methods into the growth period and the maturity period individually to increase the diversity of population by simulating different types of mutations in real plants. Inspired by various growth patterns of real plants, the Gbased growth strategy is proposed to replace a completely random growth operation of original VEGE and bias all non-optimal individuals to grow towards the current best area. We design a series of controlled experiments to evaluate the performance of our proposed strategies using 28 benchmark functions from CEC2013 suite with three different dimensions. The experimental results confirmed the mutation strategy can increase the diversity and the Gbased growth strategy plays an important role in accelerating convergence. Besides, the combination of both strategies can further improve the VEGE performance. .
6.	YU, Jun and TAKAGI, Hideyuki, Performance Analysis of Vegetation Evolution, 2019 IEEE International Conference on Systems, Man, and Cybernetics (SMC2019), 2019.10, We focus on analyzing the impact of operations of a proposed Vegetation evolution (VEGE) algorithm on its performance rather than compare it with other EC algorithms, i.e., investigate the impact of each component of the VEGE algorithm on its performance. To further analyze the performance of VEGE algorithm, we design a series of controlled experiments to investigate the contribution of each VEGE component by running them on 28 benchmark functions of 3 different dimensions. Subsequently, we summarize some our experiences on setting VEGE parameters to apply the VEGE to optimization tasks. The experimental results reveal that the maturity operation has a critical impact on performance and the number of growth operations of an individual is set as small as possible, while the number of generated seed individuals is not an important factor. Besides, we discover that population size should be gradually increased as the dimension increases. Finally, we point out several potential research directions..
7.	Yuhao Li, Jun Yu, Hideyuki Takagi, Ying Tan, Accelerating Fireworks Algorithm with Weight-based Guiding Sparks, 10th International Conference on Swarm Intelligence (ICSI2019), 2019.07, [URL].
8.	Jun Yu, Hideyuki Takagi, Ying Tan, Fireworks Algorithm for Multimodal Optimization Using a Distance-based Exclusive Strategy, 2019 IEEE Congress on Evolutionary Computation (CEC2019), 2019.06, [URL], We propose a distance-based exclusive strategy to extend fireworks algorithm as a niche method to find out multiple global/local optima. This strategy forms sub-groups consisting of a firework individual and its generated spark individuals, each sub-group is guaranteed not to search overlapped areas each other. Finally, firework individuals are expected to find different global/local optima. The proposed strategy checks the distances between a firework and other fireworks which fitness is better than that of the firework. If the distance between two firework individuals is shorter than the sum of their searching radius, i.e. amplitude of firework explosions, these two firework individuals are considered to search overlapped area. Thus, the poor firework is removed and replaced by its opposite point to track multiple optima. To evaluate the performance of our proposed strategy, enhanced fireworks algorithm (EFWA) is used as a baseline algorithm and combined with our proposal. We design a controlled experiment, and run EFWA and (EFWA + our proposal) on 8 benchmark functions from CEC 2015 test suite, that is dedicated to single objective multi-niche optimization. The experimental results confirmed that the proposed strategy can find multiple different optima in one trial run..
9.	Jun Yu, Hideyuki Takagi, Vegetation Evolution for Numerical Optimization, JPNSEC International Workshop on Evolutionary Computation, 2018.09, We propose a new population-based evolutionary algorithm (EA) by simulating vegetation growth and reproduction repeatedly to find a global optimal solution. In nature, many plants use various survival mechanism to guarantee thrive and their seeds disperse everywhere, then generated seeds root in a new suitable environment then repeat growth experience of their father. Inspired by this process, we develop a new optimization framework, where an individual consists of two periods: growth and maturity. In the growth period, each individual focus on exploitation through competing survival resources to achieve its better growth, while all individuals move to exploration by intraspecific cooperation to achieve the population continued. Therefore, the new proposal alternately performs two periods to balance exploitation and exploration. To evaluate the performance of our proposed algorithm, we compare it with the other three widely known algorithms in EA community, differential evolution (DE), particle swarm optimization (PSO) and enhanced fireworks algorithm (EFWA), and apply them to 28 benchmark functions from CEC2013 test suites of 2-dimensions (2-D), 10-D and 30-D with 30 trial runs each. The experimental results confirm that our proposal is effective and potential. Finally, we analyze the effects of the composition of our proposal on performance, and some open topics are given..
10.	Hideyuki Takagi, Keisuke Ikeda and Weiqiang Lai, Human Awareness Support by Changing Values of Hidden Factors of Input Stimuli Dynamically, 9th IEEE International Conference on Awareness Science and Technology (iCAST 2018), 2018.09, We propose an awareness support system that helps a user to be aware of the reason of his/her evaluations. Based on our proposed definition of human awareness mechanism, we extract hidden factors of input information using an auto-encoder neural networks and implement its decoder part into an awareness support system. The big feature of this system is to let a user change the values of the extracted hidden factors manually and observe the system outputs that change according to the changes of hidden factors. Experimental results using a task of generating facial emotions with 21 human subjects have shown the effectiveness of this approach..
11.	余俊, 譚営，髙木英行, Scouting Strategy for Biasing Fireworks Algorithm Search to Promising Directions, Genetic and Evolutionary Computation Conference (GECCO2018), 2018.07, We propose a scouting strategy to find better searching directions in fireworks algorithm (FWA) to enhance its exploitation capability. It generates spark individuals from a firework individual one by one by checking if the generated spark climbs up to a better direction, and this process continues until spark individual climbing down is generated, while canonical FWA generates spark individuals around a firework individual at once. We can know potential search directions from the number of consciously climbing up sparks. Besides this strategy, we use a filtering strategy for a random selection of FWA, where worse sparks are eliminated when their fitness is worse than their parents, i.e. fireworks, and become unable to survive in the next generation. We combined these strategies with the enhanced FWA (EFWA) and evaluated using 28 CEC2013 benchmark functions. Experimental results confirm that the proposed strategies are effective and show better performance in terms of convergence speed and accuracy. Finally, we analyze their applicability and provide some open topics..
12.	余俊, 譚営，髙木英行, Accelerating Fireworks Algorithm with an Estimated Convergence Point, 9th Int. Conf. on Swarm Intelligence (ICSI’2018), 2018.06, We propose an acceleration method for the fireworks algorithms which uses a convergence point for the population estimated from moving vectors between parent individuals and their sparks. To improve the accuracy of the estimated convergence point, we propose a new type of firework, the {\em synthetic firework}, to obtain the correct of the local/global optimum in its local area’s fitness landscape. The synthetic firework is calculated by the weighting moving vectors between a firework and each of its sparks and replacing the worst firework individual in the next generation. We design a controlled experiment for evaluating the proposed strategy and apply it to 20 CEC2013 benchmark functions of 2-dimensions (2-D), 10-D and 30-D with 30 trial runs each. The experimental results and the Wilcoxon signed-rank test confirm that the proposed method can significantly improve the performance of the canonical firework algorithm..
13.	Scouting Strategy Applied to Fireworks Algorithm.

2018 - 2019: Director of Fuzzy Logic Systems Institute /
2018 - 2019: Research Ethics Committee member of Fuzzy Logic Systems Institute /
2016 - 2017: Director of Fuzzy Logic Systems Institute /
2016 - 2017: Research Ethics Committee member of Fuzzy Logic Systems Institute /
2014 - 2015: Director of Fuzzy Logic Systems Institute /
2014 - 2015: Research Ethics Committee member of Fuzzy Logic Systems Institute /
2012 - 2013: Director of Fuzzy Logic Systems Institute /
2012 - 2013: Research Ethics Committee member of Fuzzy Logic Systems Institute /
2010 - 2011: Director of Fuzzy Logic Systems Institute /
2008 - 2011: Research Ethics Committee member of Fuzzy Logic Systems Institute /
2002 - 2009: Councilor of Fuzzy Logic Systems Institute /
2004: Examinor of Academis-Industry R&D Support Committee, Kyushu Industrial Technology Center /
2001 - 2003: Examinor of Academis-Industry R&D Support Committee, Institute of Systems & Information Technologies/KYUSHU /.