Adaptive dynamic programming algorithms for sequential appointment scheduling with patient preferences.

Area of Science:

• Operations Research
• Health Informatics
• Computer Science

Background:

• Effective appointment systems are crucial for optimizing medical facility utilization in outpatient departments.
• Improving efficiency and patient satisfaction are key goals for healthcare providers.

Purpose of the Study:

• To develop an advanced appointment scheduling system for outpatient departments.
• To enhance operational efficiency and patient satisfaction levels through intelligent scheduling.

Main Methods:

• A Markov decision process model was employed for sequential appointment scheduling, incorporating patient preferences.
• Adaptive dynamic programming algorithms were utilized to overcome the curse of dimensionality and improve decision-making.
• The system dynamically captures patient preferences and updates state values for optimized scheduling.

Main Results:

• Experiments demonstrated that bias-adjusted Kalman filter step-sizes yielded optimal convergence within 5000 iterations.
• A 0.9 probability of selecting a myopically optimal action optimized convergence behavior.
• Value function approximation performance varied with basis function combinations, with errors ranging from 2.7% to 8.3%.

Conclusions:

• The developed appointment scheduling system adaptively updates parameters in real-time as bookings are confirmed.
• This intelligent system significantly contributes to enhanced patient satisfaction during the appointment booking process.