Refactor Polynomial-Polynomial Multiplication in the Newton Basis

Following recent changes in the MultiIndexSet (see Issue #99 (closed)) and Grid (see Issue #135) classes, the implementation of polynomial-polynomial multiplication in the Newton basis shall be refactored.

The logic behind the multiplication is as follows: Given two polynomials in the Newton basis

• Check if the polynomials have a matching domain
• Multiply the underlying Grid instances (rely on the method of creating a product of two Grid instances; see Issue #154 (closed))
• Multiply the underlying MultiIndexSet instances if they are not the same as the one attached to the respective instances (this is related to the notion "separate indices"). Creating a product of two MultiIndexSet instances rely on the __mul__() method (see Issue #125 (closed))
• Process the coefficients; compute the values of the two polynomial operands at the unisolvent nodes associated with the product grid, multiply them to obtain the Lagrange coefficients of the product polynomial, and transform these coefficients into Newton coefficients.
• Create a new instance of polynomial with the new Grid instance, coefficients, and (when applicable) MultiIndexSet instance.

This issue is part of extending the support for arithmetic operations involving polynomial (see Issue #142 (closed)).