2D peak finding inspired by binary search

2DPeakFinding.py

def create_search_grid(center_point, side_length):
    # Return a 3x3 grid of points, centered at center_point,
    # making a square side_length tall and wide.
    return grid_of_points

def stop_condition(value):
    # Decide if you're at a good enough place to end the search.
    # Tweak this for your own application.
    return should_stop_searching

def measure_here():
    # Query the function you are trying to optimize.
    # We want to do this as few times as possible!
    return value

search_size = # Side length of approximately the area you wish to search.
              # The peak should be in here somewhere.
              
start_point = # The center point of the area to search.

resolution = # Some size below which you will stop searching.
             # Beyond this point, you just aren't going to find anything better.

best_point_and_value = (start_point, 0)

while search_size > resolution:
    for point in create_search_grid(best_point_and_value[0], search_size):
        # Go to each point and measure
        move_to(point)
        measuredValue = measure_here()
        # If that's good enough, stop
        if stop_condition(measuredValue):
            break
        # If it's the best point so far, remember it
        if measuredValue > best_point_and_value[1]:
            best_point_and_value = (point, measuredValue)
    # Search a smaller grid, centered on the best point in the current grid.
    search_size /= 2
else:
    # End the search with a final scan of size = resolution
    for point in create_search_grid(best_point_and_value[0], resolution):
        move_to(point)
        measuredValue = measure_here()
        if stop_condition(measuredValue):
            break
        if measuredValue > best_point_and_value[1]:
            best_point_and_value = (point, measuredValue)

# You found the best spot.  Go there and get on with your life.
MoveTo(best_point_and_value[0])