Drakos does not just follow the heuristics described in existing guides. Those rotations and guides are designed for a human with scarce working memory. Drakos uses dynamic programming, reinforcement learning, and Monte Carlo approaches to push the limits of efficiency.
We also work on cutting-edge statistics and mathematical optimization in our own careers. Drakos is born out of a fundamentally different approach to automation as a result. We believe in letting the math speak for itself, forgoing hand-written heuristics, which are not proven to be optimal. Instead, we treat FFXIV as a mathematical optimization problem to be solved with tools from optimization.
Most of our solvers are based on variants of DP algorithms seen in the optimization literature. Some others are based on reinforcement learning and Monte Carlo when solution time is a constraint. The challenge of Drakos is the trade-off between real solve-time and bounded optimality, something that is not often a first-class consideration in optimization research.
Our forthcoming economic engine is also based on a deep understanding of economic modeling, using auction theory and cutting-edge demand estimation frameworks to nail down market movements and deliver an intelligent system that maximizes profit on your behalf.