# A Probabilistic Guidance Approach to Swarm-to-Swarm Engagement Problem

arXiv: https://arxiv.org/abs/2012.01928 · HTML: https://mlsanity.com/paper/2012.01928/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 28 Nov 2020.

## Built on by (most cited first, 2 of 2)

- [A Scalable Reinforcement Learning Approach for Attack Allocation in Swarm to Swarm Engagement Problems](https://mlsanity.com/paper/2210.08319/) (2022)
- [Scalable Planning and Learning Framework Development for Swarm-to-Swarm Engagement Problems](https://mlsanity.com/paper/2212.02909/) (2022)

## Abstract (arXiv)

This paper introduces a probabilistic guidance approach for the swarm-to-swarm engagement problem. The idea is based on driving the controlled swarm towards an adversary swarm, where the adversary swarm aims to converge to a stationary distribution that corresponds to a defended base location. The probabilistic approach is based on designing a Markov chain for the distribution of the swarm to converge a stationary distribution. This approach is decentralized, so each agent can propagate its position independently of other agents. Our main contribution is the formulation of the swarm-to-swarm engagement as an optimization problem where the population of each swarm decays with each engagement and determining a desired distribution for the controlled swarm to converge time-varying distribution and eliminate agents of the adversary swarm until adversary swarm enters the defended base location. We demonstrate the validity of proposed approach on several swarm engagement scenarios.