Orthogonal structure on a wedge and on the boundary of a square (Q2420633)

Given three non-collinear points, a wedge is defined by fixing one point and joining it to other points by line segments. Thus a wedge is inhere given by \[ \Omega = \{ (x_{1},1): x_{1} \in [0,1]\} \cup \{ (1, x_{2}): x_{2} \in [0,1]\}, \] admitting if necessary an affine transform. Considering two nonnegative weight functions \(w_{1}\) and \(w_{2}\), defined on \([0,1]\), and the ideal \(I\) of \(\mathbb{R}[x,y]\) generated by \((1-x)(1-y)\), the bilinear form defined on \(\Omega\) by \[ \langle f,g \rangle_{w_{1} , w_{2}} = \int_{0}^{1} f(x,1)g(x,1)w_{1}(x) dx + \int_{0}^{1} f(1,y)g(1,y) w_{2}(y) dy \, , \] defines an inner product on the space of polynomials of degree at most \(n\) in two variables, modulo \(I\). The authors construct a basis of the space of orthogonal polynomials of degree \(n\) in the quotient space \(\mathbb{R}[x,y]/I\), notated by \(\mathcal{H}^{2}_{n}(w_{1},w_{2})\), for two cases, namely fixing \(w_{1}=w_{2}\) and taking Jacobi weight functions. Further, the correspondent Fourier orthogonal expansions are studied. The work developed on a wedge is then carried out for orthogonal polynomials on a parallelogram, considered, again without loss of generality, as the square \([-1,1]^{2}\). The sampling determinantal point processes associated with orthogonal polynomials are also explored.