1. The complete title of the paper published in the open literature describing the work that the author claims describes a human-competitive result: Automatic Synthesis of Swarm Behavioural Rules from their Atomic Components 2. The name, complete physical mailing address, e-mail address and phone number of each author of the paper: • Dilini Samarasinghe The School of Engineering and Information Technology The University of New South Wales at the Australian Defence Force Academy Northcott Dr, Campbell ACT 2612 d.samarasinghewidanaarachchige@student.adfa.edu.au +61 4 1503 2140 • Erandi Lakshika The School of Engineering and Information Technology The University of New South Wales at the Australian Defence Force Academy Northcott Dr, Campbell ACT 2612 e.henekankanamge@adfa.edu.au +61 2 6268 8817 • Michael Barlow The School of Engineering and Information Technology The University of New South Wales at the Australian Defence Force Academy Northcott Dr, Campbell ACT 2612 m.barlow@adfa.edu.au +61 2 6268 8955 • Kathryn Kasmarik The School of Engineering and Information Technology The University of New South Wales at the Australian Defence Force Academy Northcott Dr, Campbell ACT 2612 k.kasmarik@adfa.edu.au +61 2 6268 8023 3. The name of the corresponding author: Dilini Samarasinghe 4. The abstract of the paper: This paper presents an evolutionary computing based approach to automatically synthesise swarm behavioural rules from their atomic components, thus making a step forward in trying to mitigate human bias from the rule generation process, and leverage the full potential of swarm systems in the real world by modelling more complex behaviours. We identify four components that make-up the structure of a rule: control structures, parameters, logical/relational connectives and preliminary actions, which form the rule space for the proposed approach. A boids simulation system is employed to evaluate the approach with grammatical evolution and genetic programming techniques using the rule space determined. While statistical analysis of the results demonstrates that both methods successfully evolve desired complex behaviours from their atomic components, the grammatical evolution model shows more potential in generating complex behaviours in a modularised approach. Furthermore, an analysis of the structure of the evolved rules implies that the genetic programming approach only derives nonreusable rules composed of a group of actions that is combined to result in emergent behaviour. In contrast, the grammatical evolution approach synthesises sound and stable behavioural rules which can be extracted and reused, hence making it applicable in complex application domains where manual design is infeasible. 5. A list containing one or more of the eight letters (A, B, C, D, E, F, G, or H) that correspond to the criteria that the author claims that the work satisfies: (D) The result is publishable in its own right as a new scientific result independent of the fact that the result was mechanically created. (E) The result is equal to or better than the most recent human-created solution to a long-standing problem for which there has been a succession of increasingly better human-created solutions. (F) The result is equal to or better than a result that was considered an achievement in its field at the time it was first discovered. (G) The result solves a problem of indisputable difficulty in its field. 6. A statement stating why the result satisfies the criteria that the contestant claims: Limitations of human bias has a significant impact in leveraging the full potential of swarm systems in real world domains. As the complexity of the tasks required to be performed by the swarm system increases, mere intuition of human brain becomes insufficient to decide the rule set required for addressing the requirement. Our work proposes an evolutionary computation based approach to automatically synthesise swarm behavioural rules from their atomic components, thus making a step forward in trying to mitigate human bias from the rule generation process. (D) The result is publishable in its own right as a new scientific result independent of the fact that the result was mechanically created. In our work we propose a novel evolutionary computing based automatic synthesis mechanism for multi-agent behavioural rules where the entire rule structure can be evolved from their atomic components based on a valid syntax. Every other existing mechanism on automatic synthesis of behavioural rules require significant human involvement in designing the rule structure. The use of grammatical evolution in automatically deriving the behavioural rules entirely from scratch along with its structure, is a scientifically important result as it further mitigates human bias in the behavioural rule generation process, which is a common problem faced when designing multi-agent systems for complex task requirements. We also explore the possibility of adopting a modularised approach to generate more complex behavioural rules starting from evolved groups of simple behaviours using the proposed mechanism. Furthermore, an analysis on the rule structures generated with the model is conducted discussing the possibility of reverse engineering them for reuse in similar environments, which are all novel contributions to the field that can be published in its own right as a new scientific result. (E) The result is equal to or better than the most recent human-created solution to a long-standing problem for which there has been a succession of increasingly better human-created solutions. There exists a long history of attempts to automate the behavioural rule generation process for multi-agent systems so that human involvement can be limited and complex emerging behaviours can be achieved. The existing mechanisms often involving evolutionary approaches [1-3] and reinforcement learning [4,5], either focus on evolving parameters related to pre-designed rules or on finding the best subset from a pool of hand generated behaviours to result in required emergent behaviours. The need for rigorous manual tuning of parameters and/or the design of a pool of primary behaviours suggest that human intervention during the synthesis process is still significant. As opposed to the existing approaches, we focus on the intrinsic logic which defines the behavioural rule structure and generate rules with the atomic components (control structures, parameters, preliminary actions and logical/relational connectives) that form the said structure by mitigating human bias to a greater extent. Our comparative results with the genetic programming approach prove that the proposed grammatical evolution model is successful in evolving desired behaviours with atomic components, and it has the potential to evolve more complex behaviours in a modularised approach which is a significant achievement over the existing methods. (F) The result is equal to or better than a result that was considered an achievement in its field at the time it was first discovered. The boids model when it was first presented by Reynolds [6] was considered one of the greatest achievements in the area of simulating complex behaviours as the approach was able to generate complex emergent behaviour from only 3 simple hand crafted rules. The basic model has been used and extended in numerous researches since then, still with significant human intervention in defining the rules and parameters. We were able to achieve the same behaviours which were hand crafted by Reynolds, by only defining the atomic components and entirely automating the process of defining the rule structure, and letting the evolutionary algorithm decide the rule structure, components, parameters as well as all parameter values over the generations, thus limiting human intervention in the rule synthesis process to a huge extent than any other existing automating approach. (G) The result solves a problem of indisputable difficulty in its field. Eliminating human bias in order to generate more complex behaviours from the behavioural rule generation process for swarm and team behaviours has been a topic of discussion for decades. The existing solutions to automate the rule synthesis process have approached the problem in different perspectives including reinforcement learning, genetic algorithms and genetic programming but still possess limitations due to the significant human involvement required. We present a solution that restricts human bias in this process to a larger extent by proposing an approach that can derive entire behavioural rules from scratch by only defining their atomic components. Our work shows the potential of using grammatical evolution in evolving complex behaviours in a modularised approach with significantly less human involvement which would shape the future research in this area to focus on trying to evolve simple behaviours that can be combined to generate more complex emergent behaviours. 7. A full citation of the paper: Dilini Samarasinghe, Erandi Lakshika, Michael Barlow, and Kathryn Kasmarik. 2018. Automatic Synthesis of Swarm Behavioural Rules from their Atomic Components. In GECCO ’18: Genetic and Evolutionary Computation Conference, July 15–19, 2018, Kyoto, Japan. ACM, New York, NY, USA, Article 4, 8 pages. https://doi.org/10.1145/3205455.3205546 (To appear) 8. A statement either that “any prize money, if any, is to be divided equally among the co-authors” OR a specific percentage breakdown as to how the prize money, if any, is to be divided among the co-authors: Any prize money, if any, is to be divided equally among the co-authors 9. A statement stating why the authors expect their entry would be the “best”: The research works thus far on designing aggregates sets of behavioural rules for complex multi-agent systems have been hindered with the limitations due to human bias, and it has caused difficulty in using such systems in real world application domains. We present an evolutionary computing based approach that can remedy the issue to a greater extent than the existing approaches by handing over the entire evolution process of rule structures, components, parameters and values to the evolutionary algorithm. More importantly the generated rules can also be easily extracted and be re-used in a similar environment. This contribution would thus be valuable in future research on much complex swarm systems that can be used in real world problem domains as human involvement would only be required in identifying the atomic components relevant for the rules. Our work is a step forward in building a sophisticated world with robot groups that help with our day to day chores to defence forces consisting of swarms of drones protecting the bases. 10. An indication of the general type of genetic or evolutionary computation used: Grammatical Evolution 11. The date of publication of the paper. If the date of publication is not on or before the deadline for submission, but instead, the paper has been unconditionally accepted for publication and is “in press” by the deadline for this competition, the entry must include a copy of the documentation establishing that the paper meets the "in press" requirement.: The paper is yet to be published at the Genetic and Evolutionary Computation Conference (GECCO) July 15-19, 2018. ------------------------------------------------------ References [1] Yen-Wei Chen, Kanami Kobayashi, Hitoshi Kawabayashi, and Xinyin Huang. 2008. Application of Interactive Genetic Algorithms to Boid Model Based Artificial Fish Schools. In Knowledge-Based Intelligent Information and Engineering Systems. Springer Berlin Heidelberg, Berlin, Heidelberg, 141–148. [2] Erandi Lakshika, Michael Barlow, and Adam Easton. 2013. Co-evolving semicompetitive interactions of sheepdog herding behaviors utilizing a simple rulebased multi agent framework. In 2013 IEEE Symposium on Artificial Life (ALife). IEEE, 82–89. [3] John R Koza. 1994. Evolution of Emergent Cooperative Behavior using Genetic Programming. Computing with Biological Metaphors (1994), 280–297. [4] Adrian Agogino and Kagan Tumer. 2005. Reinforcement learning in large multiagent systems. In AAMAS-05Workshop on Coordination of Large Scale Multi-Agent Systems. Utrecht, Netherlands. [5] Yaqing Hou, Yew-Soon Ong, Liang Feng, and Jacek M Zurada. 2017. An Evolutionary Transfer Reinforcement Learning Framework for Multiagent Systems. IEEE Transactions on Evolutionary Computation 21, 4 (2017), 601–615. [6] Craig W. Reynolds. 1987. Flocks, herds and schools: A distributed behavioural model. ACM SIGGRAPH Computer Graphics 21, 4 (aug 1987), 25–34.