TA_5026201012

Summary

This presentation discusses the optimization of nursing staff scheduling in the surgery unit using simulated annealing at RSUD Sosodoro Jatikusumo, Bojonegoro. The presenter collected data and regulations for manual scheduling and identified constraints and objectives. The simulated annealing algorithm was applied to minimize the total weekly working hours violations, with parameters like alpha, initial temperature, and iterations tested for optimal solutions. The results showed no violations, and the algorithm outperformed manual scheduling in meeting hard constraints and optimizing staff scheduling.

Takeaways

• 😀 The presenter, Hilanita Sugianto, introduces their final project on optimizing nurse scheduling in the surgery unit using simulated annealing at RSUD Sosodoro Jatikusumo, Bojonegoro.
• 📊 Data collection was conducted through interviews with the head of the North Surgery Team, resulting in information on staff data and regulations used for manual scheduling.
• 🏥 The scheduling problem involves binary decision variables to determine whether a nurse is assigned to a specific shift on a given day, with constraints including daily, weekly, and monthly limits on shifts and leave.
• 🔗 Hard constraints are defined for the surgery central unit, such as daily and weekly limits on the number of shifts and specific requirements for team leaders.
• 📉 The objective function aims to minimize the number of nurses exceeding the total weekly work hours, which should be within a range of 38 to 45 hours.
• 📈 Simulated annealing was applied to optimize the scheduling for the North Surgery Team, with results showing no nurse violating the work hour limits, achieving a goal function value of zero.
• 🔧 Parameters for the simulated annealing algorithm were tested, including alpha (cooling rate), initial temperature, and number of iterations, to find the best solution.
• 📊 The results of the optimization showed that higher alpha values allowed for a slower cooling rate, leading to a broader exploration of solutions and potentially better outcomes.
• 🆚 A comparison of the optimization results with manual scheduling revealed that the manual method violated several hard constraints, while the simulated annealing and constraint programming methods did not.
• 🏅 The conclusion highlights that the simulated annealing algorithm effectively minimized the number of nurses violating work hour limits, offering a fast computation time and a solution that can be implemented with relatively minimal resources.

Q & A

What is the title of Hilanita Sugianto's final project?

The title of Hilanita Sugianto's final project is 'Optimization of Nurse Scheduling in the Operating Room Using Simulated Annealing at RSUD Sosodiro Jatikusumo, Bojonegoro'.

What type of data did Hilanita Sugianto collect for the project?

Hilanita Sugianto collected data related to the staff of the North, Central, and South surgical teams, including data on the surgical staff, HCU staff, and the regulations used for manual scheduling at the Central Surgical Unit of RSUD.

What are the hard constraints for the scheduling problem described in the project?

The hard constraints include daily limits where staff can only have one shift per day, specific rules for the head of the team needing morning shifts from Monday to Saturday except for Sundays or national holidays, and weekly limits requiring each nurse to have at least one morning, afternoon, and night shift per week.

What is the objective function of the scheduling problem in Hilanita's project?

The objective function is to minimize the total number of nurses exceeding the weekly full-time work hours range of 38 to 45 hours.

What are the results of the simulated annealing optimization for the North Surgical Team in July 2024?

The results show that there are no nurses violating the work hour range of 38 to 45 hours per week, resulting in an objective function value of zero.

What parameters were tested in the simulated annealing algorithm for the North Surgical Team?

The parameters tested were alpha (cooling rate) with values of 0.99, 0.999, and 0.995, initial temperature set at 10,000, and a fixed number of iterations at 100,000.

How does the initial temperature affect the solution quality in the simulated annealing process?

A higher initial temperature allows the algorithm to accept worse solutions initially, exploring a broader range of solutions before focusing on improving them.

What was the performance comparison between manual scheduling and simulated annealing in terms of constraint satisfaction?

Manual scheduling violated several hard constraints, while simulated annealing and constraint programming met all the constraints, indicating better performance in terms of constraint satisfaction.

What were the runtime differences between simulated annealing and constraint programming for the different teams?

For the North Surgical Team, simulated annealing took 2 minutes and 12 seconds, while constraint programming took only 2 seconds. For the HCU and South Surgical Teams, the runtime was significantly faster, ranging from 0 to 1 second for both methods.

What are the recommendations for future research based on the project findings?

Future research should consider adding scenarios related to the number of weekly work hours as soft constraints, developing datasets from other units with different regulations, and potentially integrating more complex patient integration requirements. Additionally, exploring other optimization algorithms like genetic algorithms for comparison could be beneficial.