Penyelesaian Permasalahan Penjadwalan Terintegrasi di Rumah Sakit Menggunakan Algoritma Quantum Learning - Simulated Annealing

Sitepu, Jhoni Ananta Sitepu (2026) Penyelesaian Permasalahan Penjadwalan Terintegrasi di Rumah Sakit Menggunakan Algoritma Quantum Learning - Simulated Annealing. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5026221181-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (7MB) | Request a copy

Abstract

Penjadwalan (timetabling) merupakan proses pengalokasian serangkaian acara dan sumber daya secara optimal dalam ruang dan waktu tertentu. Meskipun bertujuan memberikan solusi optimal, kegagalan dalam merancang sistem penjadwalan yang baik dapat menyebabkan inefisiensi, pemborosan sumber daya, serta ketidakpuasan pengguna layanan. Rumah sakit, sebagai institusi yang sangat bergantung pada efektivitas layanan kesehatan, menghadapi tantangan kompleks dalam mengelola jadwal operasionalnya. Penelitian ini mengangkat permasalahan integrasi tiga aspek utama penjadwalan rumah sakit yang sering kali diselesaikan secara terpisah, yaitu: penjadwalan operasi (surgical case planning), penempatan pasien ke ruang rawat inap (patient admission scheduling), dan penugasan perawat ke ruangan pasien (nurse-to-room assignment), yang berpotensi menghasilkan solusi suboptimal jika tidak ditangani secara holistik. Penelitian ini bertujuan untuk mengembangkan solusi berbasis algoritma metaheuristik, yaitu Quantum Learning - Simulated Annealing (QLSA), guna mengoptimalkan sistem penjadwalan terintegrasi rumah sakit dengan mengacu pada spesifikasi masalah Integrated Healthcare Timetabling Competition (IHTC) 2024. Data yang digunakan bersifat statis dan deterministik, dengan asumsi ketersediaan informasi di awal periode, tanpa mempertimbangkan dinamika stokastik seperti pembatalan mendadak. Metodologi penelitian menggabungkan pendekatan dua fase: fase pertama menggunakan Progressive Acceptance Iterated Local Search (PA-ILS) untuk menjamin tercapainya solusi feasible, dan fase kedua menggunakan teknik Quantum Learning untuk memandu pemilihan operator heuristik secara cerdas berbasis amplitudo kuantum guna mempercepat eksplorasi ruang solusi, yang dipadukan dengan mekanisme Simulated Annealing untuk menghindari jebakan solusi lokal (local optima) melalui kriteria penerimaan solusi probabilistik. Hasil eksperimen pada 30 instance dataset IHTC 2024 menunjukkan bahwa algoritma QLSA mampu menghasilkan solusi yang sepenuhnya feasible (layak) untuk seluruh kasus uji tanpa pelanggaran hard constraint dengan tingkat keberhasilan 100% di bawah batas waktu 10 menit. Analisis komparatif menunjukkan bahwa QLSA mengungguli algoritma dasar Hill Climbing secara signifikan dengan perbaikan nilai biaya (cost) mencapai 87,97%, serta menunjukkan stabilitas dan konsistensi yang lebih baik dibandingkan algoritma Adaptive Threshold Iterated Local Search (ATILS) dan Quantum Learning - Great Deluge (QLGD), terutama pada instance berskala industri. Dalam validasi eksternal terhadap hasil kompetisi IHTC 2024, algoritma ini menempati peringkat ke-23 secara keseluruhan dengan keunggulan nol solusi unfeasible, membuktikan reliabilitasnya dibandingkan tim-tim lain yang gagal menghasilkan solusi layak. Penelitian ini menyimpulkan bahwa penerapan algoritma QLSA terbukti efektif meningkatkan efisiensi penjadwalan rumah sakit dan menawarkan solusi yang tangguh untuk permasalahan optimasi dunia nyata yang kompleks.
===================================================================================================================================
Timetabling is the process of optimally allocating a series of events and resources within a specific space and time. Although aimed at providing optimal solutions, failure to design a good timetabling system can lead to inefficiencies, resource wastage, and service user dissatisfaction. Hospitals, as institutions heavily reliant on healthcare service effectiveness, face complex challenges in managing their operational schedules. This study addresses the complexity of integrating three main aspects of hospital timetabling often solved separately: surgical case planning, patient admission scheduling, and nurse-to-room assignment, which potentially yield suboptimal solutions if not handled holistically. This study aims to develop a solution based on a metaheuristic algorithm, Quantum Learning - Simulated Annealing (QLSA), to optimize the integrated hospital timetabling system referring to the Integrated Healthcare Timetabling Competition (IHTC) 2024 problem specifications. The data used is static and deterministic, assuming information availability at the beginning of the period, without considering stochastic dynamics like sudden cancellations. The research methodology combines a two-phase approach: the first phase uses Progressive Acceptance Iterated Local Search (PA-ILS) to ensure feasible solutions, and the second phase utilizes Quantum Learning techniques to intelligently guide heuristic operator selection based on quantum amplitude to accelerate solution space exploration, combined with the Simulated Annealing mechanism to avoid local optima traps through probabilistic solution acceptance criteria. Experimental results on 30 instances of the IHTC 2024 dataset demonstrate that the QLSA algorithm is capable of producing fully feasible solutions for all test cases without hard constraint violations, achieving a 100% success rate within the 10-minute time limit. Comparative analysis shows that QLSA significantly outperforms the baseline Hill Climbing algorithm with cost improvements reaching 87.97%, and demonstrates better stability and consistency compared to the Adaptive Threshold Iterated Local Search (ATILS) and Quantum Learning - Great Deluge (QLGD) algorithms, particularly on industrial-scale instances. In external validation against IHTC 2024 competition results, this algorithm ranks 23rd overall with the advantage of zero unfeasible solutions, proving its reliability compared to other teams that failed to produce feasible schedules. This study concludes that the application of the QLSA algorithm proves effective in improving hospital timetabling efficiency and offers a robust solution for complex real-world optimization problems.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Permasalahan Penjadwalan, Metaheuristik, Simulated Annealing, Quantum Learning, Timetabling, Metaheuristik, Simulated Annealing, Quantum Learning
Subjects:	T Technology > T Technology (General) > T57.84 Heuristic algorithms.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis
Depositing User:	Jhoni Ananta Sitepu
Date Deposited:	29 Jan 2026 07:17
Last Modified:	29 Jan 2026 07:17
URI:	http://repository.its.ac.id/id/eprint/131109

Actions (login required)

View Item