In the event of an out-of-hospital cardiac arrest (OHCA), providing emergency care within the golden rescue time is crucial for maximizing patient survival. However, due to insufficient Emergency Medical Services (EMS) resources, timely defibrillation or CPR is often not possible. To address this challenge, we leverage volunteers registered through the "First on Scene" app to provide immediate care before EMS arrival. This research aims to optimize Automated External Defibrillator (AED) placement and volunteer dispatch to maximize the expected survival rate of OHCA patients. Our approach ensures that the probability of providing emergency care within the golden rescue time meets a specified service level, considering the stochastic behavior of volunteers. We propose a stochastic simulation-optimization method using mixed binary and integer variables, incorporating data-driven models of OHCA occurrences, pedestrian flow volumes, and real road network data to simulate volunteer rescue routes. To identify the most effective AED locations and volunteer assignments, we employ a Rapid Screening algorithm based on OCBA-CO combined with a Nested Partition (NP) algorithm. In collaboration with the National Science and Technology Center for Disaster Reduction (NCDR) in Taiwan, we conduct simulations in the Sanmin District of Kaohsiung City to validate our algorithm's effectiveness and provide valuable management insights for OHCA emergency medical services.In the event of an out-of-hospital cardiac arrest (OHCA), providing emergency care within the golden rescue time is crucial for maximizing patient survival. However, due to insufficient Emergency Medical Services (EMS) resources, timely defibrillation or CPR is often not possible. To address this challenge, we leverage volunteers registered through the "First on Scene" app to provide immediate care before EMS arrival. This research aims to optimize Automated External Defibrillator (AED) placement and volunteer dispatch to maximize the expected survival rate of OHCA patients. Our approach ensures that the probability of providing emergency care within the golden rescue time meets a specified service level, considering the stochastic behavior of volunteers. We propose a stochastic simulation-optimization method using mixed binary and integer variables, incorporating data-driven models of OHCA occurrences, pedestrian flow volumes, and real road network data to simulate volunteer rescue routes. To identify the most effective AED locations and volunteer assignments, we employ a Rapid Screening algorithm based on OCBA-CO combined with a Nested Partition (NP) algorithm. In collaboration with the National Science and Technology Center for Disaster Reduction (NCDR) in Taiwan, we conduct simulations in the Sanmin District of Kaohsiung City to validate our algorithm's effectiveness and provide valuable management insights for OHCA emergency medical services.