Project Description

This project will be conducted as a two-week, on-campus summer internship with two high-school students. The work focuses on applied artificial intelligence and software development in Python through the implementation of a simulation-based environment. The proposed setting is a post-apocalyptic scenario in which multiple agents (robots or vehicles) operate from a central base. The environment includes obstacles (e.g., buildings, ruins, walls, pits) and resource locations (e.g., minerals, water, food supplies, or stranded civilians). Since the environment is partially observable, agents will explore locally, identify resources, and repeatedly transport collected resources from sites back to the base until resources are depleted. The agents aim to maximize the total collected resources within a limited time while avoiding collisions with obstacles and with each other. The environment may additionally include hazardous zones (e.g., radiation or chemical contamination/leakage) that agents may enter at a defined penalty or cost. The methodological objective is to develop a coordination mechanism inspired by swarm intelligence. In particular, agents will communicate indirectly via a pheromone-like signal deposited in the environment while operating under partial observability. The emphasis is on building a working simulation platform and evaluating baseline and learning-based strategies in a controlled setting.