Optimasi Portofolio Utang Negara Multivaluta Menggunakan Algoritma Komodo MLIPIR

R. Irfan Auliya Fauzan, Irfan (2026) Optimasi Portofolio Utang Negara Multivaluta Menggunakan Algoritma Komodo MLIPIR. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002221129_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Pengelolaan utang negara dalam berbagai mata uang asing menimbulkan tantangan berupa keseimbangan antara biaya pinjaman dan risiko pasar, khususnya risiko nilai tukar dan risiko suku bunga yang dapat meningkatkan volatilitas biaya layanan utang. Kondisi tersebut menuntut adanya strategi kuantitatif untuk menentukan komposisi portofolio utang multivaluta yang mampu meminimalkan risiko dengan tetap mempertimbangkan biaya yang diharapkan. Dalam konteks ini, optimasi portofolio utang negara yang terdiri atas Euro (EUR), Yen Jepang (JPY), dan Dolar Amerika Serikat (USD) dimodelkan menggunakan kerangka Mean-Variance Optimization (MVO). Risiko portofolio direpresentasikan oleh variansi return yang dihitung dari data historis suku bunga dan nilai tukar, sedangkan keputusan alokasi dinyatakan dalam bentuk bobot masing-masing instrumen utang. Model diformulasikan dalam dua bentuk, yaitu model skalarisasi nilai utang-risiko dan model minimum nilai utang dengan kendala risiko, serta dilengkapi kendala anggaran dan batasan bobot instrumen. Penyelesaian model dilakukan melalui dua pendekatan, yaitu fmincon sebagai solusi acuan (benchmark ) dan metode metaheuristik Komodo Mlipir Algorithm (KMA) sebagai pendekatan alternatif. Evaluasi dilakukan dengan membandingkan komposisi bobot, nilai utang, risiko, pemenuhan kendala, dan efisiensi komputasi. Hasil penelitian menunjukkan bahwa KMA mampu menghasilkan solusi portofolio utang negara multivaluta yang mendekati solusi acuan pada seluruh skenario. Pada skenario batasan umum, portofolio cenderung terkonsentrasi pada instrumen USD 30Y, sedangkan pada skenario batasan ditentukan, komposisi portofolio menjadi lebih tersebar sesuai batas minimum dan maksimum bobot. Meskipun demikian, metode fmincon memiliki waktu komputasi yang lebih rendah dibandingkan KMA. Dengan demikian, KMA dapat digunakan sebagai metode alternatif dalam optimasi portofolio utang negara multivaluta, khususnya untuk pengembangan model yang lebih kompleks.
=======================================================================================================================================
Managing sovereign debt denominated in multiple foreign currencies presents significant challenges in balancing borrowing costs with market risks, particularly exchange rate and interest rate risks, which may increase the volatility of debt servicing costs. This study aims to determine the optimal composition of a multicurrency sovereign debt portfolio that minimizes risk while considering expected borrowing costs. The optimization model considers sovereign debt denominated in the Euro (EUR), Japanese Yen (JPY), and United States Dollar (USD) using the Mean-Variance Optimization (MVO) framework. Portfolio risk is represented by the variance of returns calculated from historical interest rate and exchange rate data, while allocation decisions are expressed as the weights of each debt instrument. Two optimization models are formulated, namely a debt value-risk scalarization model and a minimum debt value model with a risk constraint, both subject to budget and instrument weight constraints. The models are solved using two approaches: the fmincon method as the benchmark solution and the Komodo Mlipir Algorithm (KMA) metaheuristic as an alternative approach. Performance evaluation is conducted by comparing portfolio weights, debt values, risk levels, constraint satisfaction, and computational efficiency. The results indicate that KMA is capable of generating multicurrency sovereign debt portfolio solutions that closely approximate the benchmark solutions across all scenarios. Under the general constraint scenario, the optimal portfolio tends to be concentrated in the USD 30Y instrument. In contrast, under the specified weight-constraint scenario, the portfolio allocation becomes more diversified while remaining within the predefined minimum and maximum weight limits. However, the fmincon method requires less computational time than KMA. Overall, the findings demonstrate that KMA is a feasible alternative for multicurrency sovereign debt portfolio optimization, particularly for future applications involving more complex optimization models.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Optimasi portofolio, Utang negara multivaluta, Mean-Variance Optimization, fmincon, Komodo Mlipir Algorithm, Portfolio optimization, Multicurrency sovereign debt.
Subjects:	H Social Sciences > HG Finance > HG4529.5 Portfolio management Q Science > Q Science (General) > Q180.55.M38 Mathematical models Q Science > Q Science (General) > Q337.3 Swarm intelligence
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	R. Irfan Auliya Fauzan
Date Deposited:	31 Jul 2026 00:55
Last Modified:	31 Jul 2026 00:55
URI:	http://repository.its.ac.id/id/eprint/140010

Actions (login required)

View Item