Skip to content

v0.40.0

Choose a tag to compare

@Torch1230 Torch1230 released this 16 Sep 14:08
· 159 commits to main since this release

CombatSolver 0.40.0

简体中文

  • “多宽度路线精炼(实验)”现在默认开启,并新增两种不同的路线筛选方式。求解器仍会先给出第一条可用路线,只有剩余时间、节点和内存充足时才继续精炼,所有搜索共用当前配置的总预算。由 0.39.0 或更早版本升级时会统一开启一次,之后仍可在性能设置中关闭。
  • 各性能档的默认节点上限提高到原来的 5 倍:低档 60,000、中档 120,000、高档 250,000、极高 500,000。自定义配置不再限制节点上限,仍需填写至少 100 的整数。
  • 新增默认关闭的“多策略路线搜索(实验)”。启用后,求解器会先用一部分现有预算探索结构不同的打法,再用剩余预算进行常规搜索,并继续按当前的战损、成长、药水和复活规则选择最终路线。该功能不会突破设置的时间和节点上限,但可能改变搜索耗时和最终结果。战斗界面会显示手动开启引导,点击后可永久隐藏。
  • 检测到皮皮极速(SpeedX)时的兼容性提醒现在可以点击并永久隐藏。

English

  • “Multi-width Route Refinement (Experimental)” is now enabled by default and includes two additional route-selection variants. The solver still publishes the first usable route before refining, and only continues when sufficient time, node, and memory budget remains. All searches share the configured total budget. Upgrading from 0.39.0 or earlier enables refinement once; it can still be disabled afterward in Performance settings.
  • Default node limits for every performance preset have been increased fivefold: 60,000 for Low, 120,000 for Medium, 250,000 for High, and 500,000 for Very High. Custom profiles no longer impose an upper node limit and still require an integer of at least 100.
  • Added an optional “Multi-strategy Route Search (Experimental)” setting, disabled by default. When enabled, the solver spends part of the existing budget exploring structurally different approaches, then uses the remaining budget for the regular search. Final routes still follow the current HP-loss, growth, potion, and revival rules. This feature stays within the configured time and node limits, but may change search time and the selected result. An in-combat banner explains where to enable it and can be dismissed permanently.
  • The SpeedX compatibility warning can now be clicked and dismissed permanently.