Movimentum - The solver algorithm can solve triangular systems

Our solver algorithm is not yet really sophisticated. Still. it can find the solutions of systems that can be solved solely by "back substitution." Here is an example of such a system:

0 = f + (e + d + c + b + a + 2.5)²
0 = e + cos(d + c + b + a + 11.5)
0 = d + (c + b + a – 0.5)²
0 = c + cos(b + a – 58)
0 = b + cos(a + 1)
0 = a + 1

In the code, I use NV to create a NamedVariable, and UE<TOp> to create a UnaryExpression with operator TOp:

    [Test]
    public void TestTriagonalSystem() {
        var constraints = new[] {
            new EqualsZeroConstraint(NV("f") +
                UE<Square>(NV("e") + NV("d") + NV("c") + NV("b") + NV("a") + new Constant(2.5))),
            new EqualsZeroConstraint(NV("e") +
                UE<Cos>(NV("d") + NV("c") + NV("b") + NV("a") + new Constant(11.5))),
            new EqualsZeroConstraint(NV("d") +
                UE<Square>(NV("c") + NV("b") + NV("a") + new Constant(-0.5))),
            new EqualsZeroConstraint(NV("c") +
                UE<Cos>(NV("b") + NV("a") + new Constant(-58))),
            new EqualsZeroConstraint(NV("b") +
                UE<Cos>(NV("a") + new Constant(1))),
            new EqualsZeroConstraint(NV("a") + new Constant(1))
        };

        IDictionary<Variable, VariableRangeRestriction> solution =
            SolverNode.Solve(constraints,
                12,
                new Dictionary<Variable, VariableRangeRestriction>(),
                0);

        AssertVariable(solution, -1, "a");
        AssertVariable(solution, -1, "b");
        AssertVariable(solution, -0.5, "c");
        AssertVariable(solution, -9, "d");
        AssertVariable(solution, -1, "e");
        AssertVariable(solution, -100, "f");
    }

Nice.