Optimization has increased its use in different domains for accurately solving challenging problems. Complex optimization problems require the use of methods that possess the capabilities to properly explore the search spaces. The traditional algorithms commonly tend to fail in suboptimal values during the optimization process; this fact affects the quality of the solutions. This situation occurs for different reasons, but the lack of diversity due to the use of exploitation operators is the most common. Brainstorm optimization is an alternative method based on the social strategy to generate new innovative ideas in work groups. In brainstorm optimization, each solution representing an idea and brainstorm process is performed using clustering algorithms. However, brainstorm optimization is not able to thoroughly explore the search space, and its diversity is reduced. It does not possess any mechanism to escape from suboptimal solutions. Besides, the computational effort is also increased in the iterative process. This paper presents a modified version of brainstorm optimization that improves its performance. In the proposed algorithm, chaotic maps and opposition-based learning are applied to initialize the solutions for a given problem. Moreover, in the optimization process, the positions of the initial population are updated using the disruptor operator. After updating the population, opposition-based learning is used again to analyze the opposite solutions. The combination of chaotic maps, opposition-based learning and disruption operator improve the exploration ability of brainstorm optimization by increasing the diversity of the population. The proposed method has been evaluated using a set of benchmark functions, and it has been also used for feature selection in data mining. The results show the high efficacy of the proposed method to determine the optimal solutions of the tested functions.