Clustering and Planning for Rescue Agent Simulation Ahmed Abouraya, Dina Helal, Fadwa Sakr, Noha Khater, Salma Osama, and Slim Abdennadher German University in Cairo, Cairo, Egypt {dina.helal,slim.abdennadher}@guc.edu.eg Abstract. The paper describes the contribution of the GUC ArtSapience team to the Rescue Agent Simulation competition in RoboCup in terms of the current research approach. The approach is divided into two parts: clustering and planning. Clustering is done through task allocation to di- vide the map among the agents. Planning is done after assigning the agents to parts of the map to determine how they should cooperate and coordi- nate together and how they should prioritize their tasks [2]. The agents can coordinate together using centers and communication if available or dynamically without the use of communication. 1 Introduction Rescue planning and optimization is one of the emerging fields in Artificial Intel- ligence (AI) and Multi-Agent Systems. The RoboCup Rescue Agent Simulation provides an interesting test bench for many algorithms and techniques in this field. The simulation environment provides challenging problems that combine optimization (routing, planning, scheduling) and multi-agent systems (coordina- tion, communication, noisy or missing communication) [3]. The Robotics and Multi-Agent Systems (RMAS) research group at the Ger- man University in Cairo (GUC) was established in September 2010. The goal of the research group is to study and develop AI algorithms to solve problems in robotics and simulation systems. These fields include computational intelligence, computer vision, multi-agent systems, and classical AI approaches. The GUC ArtSapience team made its third participation to the Rescue Agent Simulation in 2013 and won the first place. Our first participation (as RMAS ArtSapience) was in 2011 and the team ranked third in the final round. Our second participation was in 2012. In this paper, we propose our approach which is divided into two parts: cluster- ing and planning. Clustering is done using K-means to divide the map into small regions. The number of clusters is relative to the number of agents in the map. This is done in the preprocessing phase to set the initial positions of the agents in the map. This ensures that all regions in the map are covered by the three types of agents. As the simulation starts, each agent has a list of tasks to do which are prioritized. Agents communicate and coordinate together using communication channels and voice messages. S. Behnke et al. (Eds.): RoboCup 2013, LNAI 8371, pp. 125–134, 2014. c  Springer-Verlag Berlin Heidelberg 2014