Modeling Collective Behaviors: A Moment-Based Approach

In this work we introduce an approach for modeling and analyzing collective behavior of a group agents using moments. We represent the via their distribution derive method to estimate dynamics use predict evolution by first computing moment trajectories then reconstruct agents. latter inverse problem is solved in order nominal recover macro-scale properties The proposed applicable several types multi-agent systems, e.g., leader-follower systems. error bounds describe how take these into account dynamics. convergence also analyzed cases with monomial To illustrate theory, two numerical examples are given. consider system interactions compare methods second example apply framework pedestrian crowd