Replace trove4j dependency with fastutil

[rolnico]

Caution

Not for the 06/2026 release

Please check if the PR fulfills these requirements

• The commit message follows our guidelines
• Tests for the changes have been added (for bug fixes / features)
• Docs have been added / updated (for bug fixes / features)
• A PR or issue has been opened in all impacted repositories (if any)

Does this PR already have an issue describing the problem?
Follows powsybl/powsybl-core#3580

What kind of change does this PR introduce?
Quality

What is the current behavior?
OLF uses trove4j which is abandonned

What is the new behavior (if this is a feature change)?
OLF now uses fastutil

Does this PR introduce a breaking change or deprecate an API?

• Yes
• No

If yes, please check if the following requirements are fulfilled

• The Breaking Change or Deprecated label has been added
• The migration steps are described in the following section

What changes might users need to make in their application due to this PR? (migration steps)

Other information:

[rolnico]

Benchmarks comparison done using powsybl-benchmark for the version using Trove4J and the one using FastUtil:

Load flow benchmark

Network	Basic	Δ	Standard	Δ	Standard (no reactive limits)	Δ
IEEE 14	162 µs → 143 µs	-11.7%	155 µs → 157 µs	+1.3%	149 µs → 156 µs	+4.7%
IEEE 118	1.10 ms → 1.16 ms	+5.5%	1.48 ms → 1.53 ms	+3.4%	1.19 ms → 1.20 ms	+0.8%
IEEE 300	2.61 ms → 2.76 ms	+5.7%	4.42 ms → 4.53 ms	+2.5%	3.43 ms → 3.15 ms	-8.2%
RTE 1888	21.3 ms → 20.7 ms	-2.8%	24.5 ms → 25.2 ms	+2.9%	23.1 ms → 23.4 ms	+1.3%
RTE 6515	90.6 ms → 96.5 ms	+6.5%	143 ms → 149 ms	+4.2%	112 ms → 101 ms	-9.8%
RealGrid	99.7 ms → 96.5 ms	-3.2%	153 ms → 146 ms	-4.6%	121 ms → 106 ms	-12.4%

Mono-thread security analysis benchmark

Network	Basic	Δ	Standard	Δ	Standard (no reactive limits)	Δ
IEEE 14	46 µs → 48 µs	+4.3%	74 µs → 74 µs	0.0%	55 µs → 53 µs	-3.6%
IEEE 118	198 µs → 199 µs	+0.5%	422 µs → 415 µs	-1.7%	263 µs → 268 µs	+1.9%
IEEE 300	837 µs → 727 µs	-13.1%	1.93 ms → 1.59 ms	-17.6%	1.06 ms → 0.927 ms	-12.5%
RTE 1888	4.2 ms → 4.0 ms	-4.8%	6.9 ms → 5.7 ms	-17.4%	5.1 ms → 4.9 ms	-3.9%
RTE 6515	17.3 ms → 16.5 ms	-4.6%	19.5 ms → 19.1 ms	-2.1%	19.3 ms → 16.0 ms	-17.1%

Sensitivity analysis benchmark

Network	Basic	Δ	Standard	Δ	Standard (no reactive limits)	Δ
IEEE 14	67 µs → 65 µs	-3.0%	95.5 µs → 97 µs	+1.6%	74.4 µs → 74 µs	-0.5%
IEEE 118	3.3 ms → 3.3 ms	0.0%	3.7 ms → 3.6 ms	-2.7%	3.4 ms → 3.4 ms	0.0%
IEEE 300	4.2 ms → 3.9 ms	-7.1%	5.6 ms → 5.0 ms	-10.7%	4.1 ms → 4.2 ms	+2.4%
RTE 1888	9.4 ms → 8.9 ms	-5.3%	12.1 ms → 11.0 ms	-9.1%	10.1 ms → 10.1 ms	0.0%
RTE 6515	22.9 ms → 24.4 ms	+6.6%	39.2 ms → 26.2 ms	-33.2%	27.4 ms → 22.2 ms	-19.0%
RealGrid	26.8 ms → 23.7 ms	-11.6%	30.9 ms → 25.3 ms	-18.1%	21.9 ms → 26.1 ms	+19.2%

The base branch was changed.