RAHARJO, Agustinus and Isnanto, R. Rizal and Surarso, Bayu (2026) Komparasi Kinerja CP-SAT OR-Tools, Tabu Search, dan Hybrid Genetic Algorithm-Simulated Annealing pada Nurse Rostering Problem dengan Kendala Shift Malam dan Team Coverage. Masters thesis, UNIVERSITAS DIPONEGORO.

	Text 1. Cover awal.pdf Download (161kB)
	Text 2. Cover lengkap.pdf Restricted to Repository staff only Download (1MB)
	Text 3. BAB I.pdf Download (323kB)
	Text 4. BAB II.pdf Download (710kB)
	Text 5. BAB III.pdf Restricted to Repository staff only Download (661kB)
	Text 6. BAB IV.pdf Restricted to Repository staff only Download (1MB)
	Text 7. BAB V.pdf Restricted to Repository staff only Download (312kB)
	Text Daftar Pustaka.pdf Download (286kB)
	Text Lampiran.pdf Restricted to Repository staff only Download (698kB)

Abstract

Penjadwalan perawat (Nurse Rostering Problem/NRP) penting bagi kesinambungan layanan, beban kerja, dan keselamatan pasien. Di banyak rumah sakit, penjadwalan masih dilakukan manual sehingga menyita waktu, rawan kesalahan, dan sulit menjaga kepatuhan aturan, terutama struktur shift malam dan team coverage. Penelitian ini memformalkan model NRP yang menekankan dua kendala tersebut dan membandingkan tiga pendekatan optimasi: solver CP-SAT pada Google OR-Tools sebagai pendekatan eksak, Tabu Search sebagai metaheuristik, dan Hybrid Genetic Algorithm–Simulated Annealing (HGASA) sebagai metode hibrida. Pengujian dilakukan pada skenario penjadwalan bulanan skala kecil, menengah, dan besar dengan variasi batas waktu komputasi. Kinerja dievaluasi menggunakan metrik kelayakan solusi, jumlah pelanggaran kendala, pemerataan beban kerja, pemenuhan kebutuhan shift, dan waktu komputasi. Kontribusi utama penelitian ini adalah menyajikan perbandingan lintas pendekatan pada konfigurasi kendala yang sama disertai analisis sensitivitas terhadap skala dan batas waktu, sehingga menghasilkan rekomendasi pemilihan algoritma yang lebih operasional. Data dan skenario berasal dari studi kasus SMC RS Telogorejo Semarang serta variasi data dummy. Hasil menunjukkan CP-SAT paling konsisten menghasilkan jadwal layak dengan pelanggaran minimal dan pemerataan beban kerja yang baik; HGASA kompetitif pada skala kecil namun lebih sensitif saat skala meningkat; sedangkan Tabu Search menurun pada skala besar. Rentang time limit 90–120 detik memberikan kompromi terbaik antara kualitas jadwal dan waktu tunggu pengguna. Prototipe modul penjadwalan mendukung otomasi penyusunan jadwal dan alokasi shift malam serta team coverage yang lebih adil.
Kata-kunci: Nurse Rostering Problem, CP-SAT, Tabu Search, Hybrid Genetic Algorithm–Simulated Annealing (HGASA)

Nurse rostering (Nurse Rostering Problem/NRP) is critical for continuity of care, nurses’ workload, and patient safety. In many hospitals, rostering is still done manually, making the process time-consuming, error-prone, and difficult to keep compliant with work rules, especially night-shift structure and team coverage. This study formalizes an NRP model that emphasizes these two constraints and compares three optimization approaches: CP-SAT (OR-Tools) as an exact method, Tabu Search as a metaheuristic, and Hybrid Genetic Algorithm–Simulated Annealing (HGASA) as a hybrid method. Experiments are conducted on monthly rostering scenarios of small, medium, and large scale under varying computation time limits. Performance is evaluated using solution feasibility, the number of constraint violations, workload balance, shift coverage fulfillment, and computation time. The main contribution of this study is a cross-approach comparison under the same constraint configuration (night-shift structure and team coverage), accompanied by sensitivity analysis on problem scale and time limits, producing more operationally relevant guidance for algorithm selection. Data and scenarios are derived from a case study at SMC RS Telogorejo Semarang and supplemented with dummy data variations. Results show that CP-SAT is the most consistent in producing feasible schedules with minimal violations and good workload balance; HGASA is competitive at small scale but more sensitive as scale increases; while Tabu Search degrades on larger instances. A time limit of 90–120 seconds provides the best trade-off between schedule quality and user waiting time. The prototype scheduling module supports rostering automation and promotes fairer allocation of night shifts and team coverage.
Keywords: Nurse Rostering Problem, CP-SAT, Tabu Search, Hybrid Genetic Algorithm–Simulated Annealing (HGASA)

Actions (login required)

View Item