Modeling multi-scale complex systems is challenging due to the coupled dynamics across various scales of the systems. For these systems, the critical problem lies in devising optimal sampling strategies that effectively capture these dynamics without incurring unnecessary computational costs or missing crucial details. Current state-of-the-art methods often fail to adapt dynamically to multi-scale complexities, leading to inefficiencies and inaccuracies. Furthermore, these existing strategies lack robustness in noisy real-life systems and have limitations in scalability when confronting high-dimensional systems. To innovatively address the optimal sampling in the multi-scale system discovery, our approach hinges on integrating a tailored reinforcement learning framework inspired by deep -learning with a system discovery framework—Sparse Identification of Nonlinear Dynamical Systems (SINDy). Rather than a generic reward system, we crafted an adaptive reward signaling mechanism that dynamically resonates with the system's intricate dynamics. This innovation ensures the agent doesn't just collect data but intelligently adapts its sampling in response to shifting multi-scale patterns. This adaptability is crucial in ensuring robustness against noise and scalability to high-dimensional systems. The performance and effectiveness of our method were assessed through two rigorous numerical studies: (1) a coupled noisy fast and slow Van der Pol oscillator, and (2) a noisy fast Van der Pol oscillator coupled with a noisy slow Lorenz system. In each case, the reinforcement learning model demonstrated its robustness and efficiency by autonomously determining the optimal data sampling strategy to capture the multi-scale dynamics accurately. In addition to those numerical studies, we incorporated the Wendling neural mass model, further challenging our RL model with its intricate neural dynamics. We observed that the reinforcement learning model in those scenerios could develop intricate policies that significantly mitigated non-convergence issues, reduced the sample size, and enhanced the robustness of SINDy in the presence of noise. Our research significantly advances data-efficient reinforcement learning for multi-scale complex systems, enhancing system understanding, prediction, and control, and opening new avenues for optimized sampling and precise system discovery. 

In organizational and academic settings, the strategic formation of teams is paramount, necessitating an approach that transcends conventional methodologies. This study introduces a novel application of multi-criteria integer programming (MCIP), which simultaneously accommodates multiple criteria, thereby innovatively addressing the complex task of team formation. Unlike traditional single-objective optimization methods, our research designs a comprehensive framework capable of modeling a wide array of factors, including skill levels, backgrounds, and personality traits. The objective function of this framework is optimized to maximize within-team diversity, while minimizing both conflict levels and variance in diversity between teams. Central to our approach is a two-stage optimization process. Initially, it segments the population into sub-groups using a weighted heterogeneous multivariate K-means algorithm, allowing for a targeted and nuanced team assembly. This is followed by the application of a surrogate optimization technique within these sub-groups, efficiently navigating the complexities of MCIP for large-scale applications. Our approach is further enhanced by the inclusion of explicit constraints such as potential interpersonal conflicts, a factor often overlooked in previous studies. The results from our study demonstrate the optimality robustness of our model across simulation scenarios with different data heterogeneity levels. The contributions of this study are manifold, addressing critical gaps in existing literature with a theory-backed, empirically validated framework for advanced team formation. Beyond theoretical implications, our work provides a practical guide for implementing conflict-aware, sophisticated team formation strategies in real-world scenarios. This advancement paves the way for future research to explore and enhance this model, providing more sophisticated and efficient team formation strategies.