No abstract

Many real-world problems involve cooperation and/or competition among multiple agents. These problems often can be formulated as multi-agent problems. Recently, Reinforcement Learning (RL) has made significant progress on single-agent problems. However, multi-agent problems still cannot be easily solved by traditional RL algorithms. First, the multi-agent environment is considered as a non-stationary system. Second, most multi-agent environments only provide a shared team reward as feedback. As a result, agents may not be able to learn proper cooperative or competitive behaviors by traditional RL. Our algorithm adopts Evolution Strategies (ES) for optimizing policy which is used to control agents and a value decomposition method for estimating proper fitness for each policy. Evolutionary Algorithm is considered as a promising alternative for signal-agent problems. Owing to its simplicity, scalability, and efficiency on zeroth-order optimization, EAs can even outperform RLs on some tasks. In order to solve multi-agent problems by EA, a value decomposition method is used to decompose the team reward. Our method is parallel on multiple cores, which can speed up our algorithm significantly. We test our algorithm on two benchmarking environments, and the experiment results show that our algorithm is better than traditional RL and other representative gradient-free methods.

Network representation learning (NRL) has been widely used to help analyze large-scale networks through mapping original networks into a low-dimensional vector space. However, existing NRL methods ignore the impact of properties of relations on the object relevance in heterogeneous information networks (HINs). To tackle this issue, this paper proposes a new NRL framework, called Event2vec, for HINs to consider both quantities and properties of relations during the representation learning process. Speci cally, an event (i.e., a complete semantic unit) is used to represent the relation among multiple objects, and both event-driven rst-order and second-order proximities are de ned to measure the object relevance according to the quantities and properties of relations. We theoretically prove how event-driven proximities can be preserved in the embedding space by Event2vec, which utilizes event embeddings to facilitate learning the object embeddings. Experimental studies demonstrate the advantages of Event2vec over state-ofthe-art algorithms on four real-world datasets and three network analysis tasks (including network reconstruction, link prediction, and node classi cation).

No abstract

No abstract

Brain storm optimization algorithms (BSO) have shown great potential in many global black-box optimization problems. However, the existing BSO variants can suffer from three problems: (1) large-scale optimization problem; (2) hyperparameter optimization problem; (3) high computational cost of the clustering operations. To address these problems, in this paper, we propose a simple yet effective BSO variant named Brain Storm Optimization Algorithm with Cooperative Learning Strategy (BSO-CLS). It is inspired by the new ideas generating process of brain storm in which the participators propose their own ideas by cooperatively learning other participators’ ideas. Thus, BSO-CLS iteratively updates the candidate solutions by linearly combining other solutions with the weights deriving from the fitness values of other solutions. To validate the effectiveness of the proposed method, we test it on 6 benchmark functions with the 1000 dimensions. The experimental results show that BSO-CLS can outperform the vanilla BSO and the other BSO variant with the learning strategy.

No abstract

No abstract