My approach is of complexity O (kn), where k = range (min (array) to max (array)) (inclusive) and n = size of the array. It passed all the test cases.
REMEMBER: The point of convergence doesn't HAVE to be an element from the array. It can be any number in the range of min to max, so I applied exhaustive search.

def get_optimal (array, size, pivot, op1, op2):
inc = dec = cost = 0
for i in array:
if (i < pivot):
inc += (pivot - i);
elif (i > pivot):
dec += (i - pivot);

if (dec > inc):  
    return (float ('inf'));  

cost += dec * op1;  
inc -= dec;  
cost += (inc * op2);  
return (cost);  

size, op1, op2 = [int (i) for i in input ().split ()];
array, optimal = [int (i) for i in input ().split ()], float ('inf');

for i in range (min (array), max (array) + 1):
temp = get_optimal (array, size, i, op1, op2);
optimal = temp if optimal > temp else optimal;
print (optimal);