Improve Documentation about Memory Access Cost

[FloydZ]

Currently many people have the following mental model for memory access cost: T + \log(M) (log memory costs). But in reality it looks more like this:

    def _compute_estimate_time_complexity(self, params):
        """Computes time complexity in estimate mode and converts to bit complexity.
    
        Computes time complexity (in estimate mode) for given parameter set and converts to bit complexity
        and adds memory access costs if set.
        """
        temp_time_complexity = self._compute_time_complexity(params)
        if self.bit_complexities:
            temp_basic_operation_cost = self.problem.to_bitcomplexity_time(
                temp_time_complexity)

            if self._memory_access != 0:
                temp_memory_access_cost = temp_time_complexity
                temp_memory_access_cost += self.memory_access_cost(
                    self.memory_complexity())
                temp_time_complexity = log2(int(2 ** temp_basic_operation_cost)
                                            + int(2 ** temp_memory_access_cost))
            else:
                temp_time_complexity = temp_basic_operation_cost

        return temp_time_complexity