Sparse grids are carefully selected subgrids of full grids on hypercubes, enabling the construction of high-accuracy approximations with significantly fewer grid points. Their efficiency makes them particularly suitable for high-dimensional applications. Constructed hierarchically using the combination technique, sparse grids blend smaller tensor product grids, resulting in operators that are combinations of tensor product operators. This structure effectively mitigates the curse of dimensionality and facilitates efficient computations in high-dimensional settings. In this paper, we present numerical results for the approximation and optimization of functions on sparse grids using spline-based interpolation and quasi-interpolation. Our findings validate well-established theoretical results, demonstrating that sparse grid operators achieve performance comparable to full grid operators while requiring substantially less storage.