branchbefore/main/GoalImprovementscancel
6 Commits over 0 Days - ∞cph!
Fixed minor bug making things not work as intended.
Switched from using a ceiling threshold of previous best decisions this tick to using a previous best decisions memory array, that we make sure we evaluate everything we need, even in cases where score at the same level in the goal hierarchy might score the same where further down into the hierarchy they will score differently. First come, first served.
Fixed recursive decision making over goals, goal plans and goal plan variants, that we exhaust all our options rather than to not reach a decision.
Fixed some issues with regards to considerations being non-existen or there only being a single consideration in our DSE.
Cleaned up debug in DSE a bit, that it's not as intrusive on the code elegance.
Made Agent+DM a bit more robust, where if finding the best goal's goal plan or goal plan variant doesn't pull through and fail, it will go on and try to find the next best goal.
