Designing novel algorithms is very hard. You can borrow some ideas from the existing literature, but you have to figure out how to apply them to your problem a

Designing novel algorithms is very hard. You can borrow some ideas from the existing literature, but you have to figure out how to apply them to your problem and that can take some nontrivial inspiration. I have one paper with a few new algorithms in it and that took me several years to write in grad school. The algorithms themselves aren't that complicated, but the concepts and theorems we had to come up with to support them required some genuine insight.