diff --git a/src/rewards.jl b/src/rewards.jl index 4700c5e..948c395 100644 --- a/src/rewards.jl +++ b/src/rewards.jl @@ -2,7 +2,7 @@ function POMDPs.reward(p::BoxWorld, s::State, a::MoveAction) box = s.pos boxp = p.boxes[a.target].pos - return euclidean(box, boxp) * -1. + return -1. * euclidean(box, boxp) end function POMDPs.reward(p::BoxWorld, s::State, a::TakeAction)