This scheduler is capable of determining optimal strategies for deploying patrols around the UPLB campus. Deployments are effective in making sure that high-priority areas get an adequate amount of visits, while lower priority areas are not left completely vulnerable. Implementing randomization helps in avoiding predictability. Adversaries would find it harder to find attack opportunities as the patrols follow ever changing patterns. Overall, this system can make patrols more effective without increasing the cost.
The current implementation of the program has several limitations. The main targets for improvement would be in the implementations for zoning, scheduling and patrol allocations. A more dynamic approach would enable more dynamic and adaptive solutions. One possible solution for the zoning programs is to dynamically divide the campus according to the number of patrols. The divisions would need to be done within the program, but theoretically this would maximize the spread of patrols, which reduces the response time.
If possible, the simulation should be integrated into the program itself. Having to run the scheduler and the simulation programs separately would mean that data sharing between the two would be prone to errors, human error in particular. If the simulation program is successfully integrated, it could also provide the patrol agents' location data in real-time to the scheduler. Given the said data, the scheduling program can perform more accurate computations, particularly in patrol mileages.
After further refinements, future studies may adapt the scheduling algorithm for use in other areas facing urban crime. There would be distinct differences between domains, but the principle of weighted probabilities and mixed strategies remain the same. Future studies may work on larger, more open domains, such as subdivisions, or in public events and conventions.