This week, our team shifted our focus toward prototype development following the feedback we received during our Preliminary Design Review at CAE. We began working on a simplified simulation involving a small number of agents to establish a baseline for measuring performance. Our initial approach focuses on implementing a brute-force algorithm where every agent interacts with every other agent, resulting in O(n^2) complexity. From there, we plan to explore spatial partitioning techniques to reduce unnecessary interactions and improve scalability as we increase the number of agents. We also continued researching simulation frameworks and discussing how to handle entity generation and visualization without taking too much time away from algorithm development. During our meetings, we discussed keeping the initial simulation simple, potentially using a red-versus-blue aircraft scenario, while working toward larger-scale simulations and GPU acceleration. With the prototype day approaching, our main priority is building a working foundation we can build on and use to evaluate performance improvements.

This week, our team placed the finishing touches on our Preliminary Design Review before visiting CAE USA headquarters. On October 15th we headed over to Tampa to meet with our liaison Brian Stensrud and several other senior technical employees at CAE. Our presentation went well and we received lots of great feedback and advice from those in attendance. After our presentation and Q&A we received a presentation on career opportunities from CAE’s HR team and then a tour of a C-130 simulator facility on CAE’s campus. Overall the trip was very successful. Next week we will look to regroup based on the CAE team members’ feedback and look ahead to completing our prototype.

This week, our team continued preparing for our upcoming Preliminary Design Review at CAE USA. We focused on finalizing our research into different algorithms and techniques that could improve the performance and scalability of our simulation system. We also worked on comparing approaches such as spatial partitioning, scheduling, and GPU acceleration to determine which areas would be most beneficial to explore further. Alongside our research, we discussed potential simulation frameworks and how we could use existing tools to simplify parts of our prototype development. We spent the rest of the week organizing our findings, refining our project goals, and preparing to present our proposed approaches to CAE. Next week, we look forward to visiting CAE’S headquarters in Tampa, receiving feedback on our research, and using that feedback to guide the development of our prototype.

This week, our team focused on preparing and presenting our rough draft for the Preliminary Design Review (PDR). We brought together the research and planning we have completed so far, including our project requirements, technical performance measures, preliminary architecture, and the concepts we are considering for the prototype. The presentation gave us a chance to organize our work into a clearer overall direction and receive feedback before the final PDR. We also continued to discuss approaches such as spatial interaction management and GPU-based execution, but we have not begun implementing these concepts yet. Our next steps will be to use feedback from the PDR process to narrow our approach and begin moving toward prototype development.

This week, our team focused on researching approaches for improving the scalability of large constructive simulations. We investigated algorithms including flow-field pathfinding, GPU-parallel A*, hierarchical pathfinding, and GPU-based BVH/LBVH structures for handling entity movement, spatial queries, and interactions more efficiently. We also researched existing constructive simulation systems such as OneSAF to better understand how large simulations organize entities, behaviors, and interactions. Based on discussions with our CAE liaison, our next step is to narrow this research into a few promising approaches and explore small implementations or demos to help evaluate their performance and scalability.

This week, our team made progress in defining the technical direction of the project. After meeting with our CAE liaison and coach, we gained a clearer understanding of the project requirements and determined that performance and scalability will be the primary focus of our simulation. We divided our research effort between existing constructive simulation systems and scalable entity-simulation algorithms, while also exploring technologies such as CUDA for future GPU implementation. In addition, we began refining our Product Design Specifications (PDS) and discussing possible simulation domains, entity types, and interaction models.

Welcome to our Massively Multiplayer Simulation project sponsored by CAE USA. Our team is working to develop a constructive simulation that can support large numbers of interacting computer-generated entities while improving scalability compared to traditional approaches. Throughout the project, we will explore scalable interaction methods, constructive simulation techniques, and GPU computing technologies. This blog will document our team’s progress, technical decisions, challenges, and development throughout the year.