Prioritized Alpha Effect

Problem

Compute the sampling-probability ratio induced by the prioritized replay exponent alpha.

Return (p_max / p_min)^alpha. If p_min is not positive or p_max is below p_min, return 0.0.

Hint

Alpha of 0 flattens the ratio to 1 and alpha of 1 passes the raw priority ratio through.

case 1

in: [1,100,0.5]

expected: 10

case 2

in: [1,10,2]

expected: 100

case 3

in: [2,8,1]

expected: 4

case 4

in: [0,5,1]

expected: 0

case 5

in: [5,5,0.7]

expected: 1

Starter codePython

def prioritized_alpha_effect(p_min, p_max, alpha):
    # Your code here
    pass

5 test cases · run your code against them in the editor