A Personalised Optimising Level Adaptation (OLA) Difficulty Algorithm for Scenario Simulations in Professional VR Simulators

Aim: This study introduces the Optimising Level Adaptation (OLA) algorithm, designed to enhance scenario simulations for professional VR training by dynamically adjusting difficulty levels to match user performance, thereby supporting personalised learning and readiness for high-stakes situations such as firefighting and emergency response. Project and methods: The OLA algorithm divides scenario activities into blocks and adjusts their difficulty based on user performance in comparison to a reference group of AI-controlled agents. The algorithm’s efficacy was tested across three proprietary VR simulators covering diverse professional scenarios: public speaking, hydrogen electrolysis and mechanical technician operations. Each scenario was divided into ten blocks of varying difficulty (easy, medium, difficult), dynamically adjusted based on the user's performance. This structure enables rapid adaptation, making it particularly beneficial for fire and rescue training, where realistic, yet scalable, scenario complexity is critical to preparing for unpredictable conditions in the field. Results: Testing with 30 participants per simulator revealed an average final score of approximately 75%, closely aligning with the target success rate of 70%. The average number of difficulty level switches (between 0.8 and 1.16 across scenarios) demonstrated the algorithm’s effective adaptation to user performance, thus ensuring optimal engagement. The OLA algorithm’s capacity to tailor training difficulty in real time reflects its potential to enhance skill retention and readiness in emergency response settings, where maintaining user engagement at appropriate challenge levels is essential for preparedness in life-threatening situations. Conclusions: The OLA algorithm provides significant advancements in personalised VR training, particularly within fire- and rescue-related applications, by maintaining optimal engagement and adaptive challenge levels. The adaptability demonstrated across multiple scenarios indicates its versatility and potential for use in diverse high-risk training applications. Future research could enhance the OLA algorithm’s effectiveness by refining scenario block determination, therefore contributing to improved response times, decision-making and operational efficiency in the emergency services. Keywords: VR training, personalised learning, adaptive difficulty, fire and rescue training, scenario simulation, professional development, Optimising Level Adaptation (OLA), emergency response preparedness