Loading...

Application of Data Mining Methods in Operation Theater Scheduling for Elective Patients

Ahmadian, Mohammad Amin | 2022

93 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 55373 (01)
  4. University: Sharif University of Technology
  5. Department: Industrial Engineering
  6. Advisor(s): Varmazyar, Mohsen
  7. Abstract:
  8. According to the growing population and the need to make optimal use of available resources, the discussion of planning and scheduling in healthcare will be useful. One of the most expensive and lucrative sections in a hospital, which always spends a large percentage of renewable and non-renewable resources, is the Operating rooms (ORs). These ORs are the most important part of the hospital in terms of health and financial issues. In these scheduling problems, in order to be able to model the real situation, we needed to consider the recovery section integrated with the ORs. Also, the duration of surgery and stay in the recovery rooms are considered stochastic. In this study, we solve the problem of an ORs scheduling which has 3 operating rooms and 6 recovery beds. Simulation is commonly used in the literature to solve such problems and estimate the duration of surgical operations and then solve the scheduling problems. We use an artificial neural network (ANN) to properly estimate the values of the objective functions of the sequences. These ANNs can, at the appropriate time, estimate our response with significant error. In this model, we first generate the basic data required for ANN training in random sequences and then evaluate them by simulating them. Then, the neural network is trained by the previous data. Now, by the ANNs, we can predict the amount of problem’s objective function, for each number of patients with whom we trained the ANN. Population-based algorithms can be used to solve problems such as our input being a set of answers. To solve this problem, we have used genetic algorithms (GA) and particle swarm optimization algorithms (PSO). In this research, in order to estimate the better answer by these algorithms, the Taguchi method has been used to design experiments by which the parameters of meta-heuristic algorithms can be estimated. In these algorithms, there needs to be a fitness function to measure the answers obtained in each period, so that the algorithm in each iteration gives us an optimal answer. We used the artificial neural network we taught earlier for this purpose. Finally, we evaluate the answers and results obtained from these two algorithms. Comparing the two meta-heuristic algorithms used, we conclude that the genetic algorithm gives a better answer than the particle swarm optimization algorithm.
  9. Keywords:
  10. Scheduling ; Operating room ; Data Mining ; Artificial Neural Network ; Genetic Algorithm ; Particles Swarm Optimization (PSO)

 Digital Object List

 Bookmark

No TOC