ECON10071 Chap.12 Definiteness and Concave Programming
Definiteness and Concave Programming
Definiteness and Concave Programming connect matrix signs to optimisation curvature. A symmetric matrix defines a quadratic form, with each cross-term coefficient split across two equal off-diagonal entries. Leading principal minors identify positive and negative definiteness. When a Hessian is negative definite, the objective is strictly concave.
On a convex feasible set, a valid Kuhn-Tucker point is then a unique global maximum when the stated regularity conditions hold.
What this chapter covers
- 01
Symmetric quadratic forms
- 02
Cross-term coefficients
- 03
Leading principal minors
- 04
Positive definiteness
- 05
Negative definiteness
- 06
Concave-program sufficiency
Worked example · free
Classify a Hessian matrix
- 3The first leading minor is -4 and the determinant is 7, giving the negative-then-positive pattern.
- 3H is negative definite, so an objective with this constant Hessian is strictly concave and has at most one maximiser on a convex set.
Key terms
- Quadratic form
- A scalar expression written as x transpose A x.
- Principal minor
- A determinant formed from matched leading rows and columns for the standard sign test.
- Strict concavity
- Curvature that makes any feasible stationary maximum unique on a convex set.
Definiteness and Concave Programming FAQ
How is a cross term placed in a symmetric matrix?
A term 2kxy is generated by the two equal products a12xy and a21yx, so each off-diagonal entry is k. Forgetting the split doubles the cross effect and changes the determinant.
What pattern proves negative definiteness in two variables?
The first leading principal minor must be negative and the determinant positive. That alternating sign pattern makes the quadratic form negative for every nonzero vector.
Why does concavity matter for Kuhn-Tucker conditions?
Concavity of the objective and convexity of the feasible set rule out better separated local peaks. Under the stated conditions, a feasible Kuhn-Tucker candidate is globally optimal, and strict concavity makes it unique.
Exam move
Translate each quadratic expression into a symmetric matrix before applying sign tests. Compute leading principal minors without rounding and classify strict, semidefinite or indefinite cases carefully. Then connect the matrix result to convexity or concavity and to the strength of an optimisation conclusion. Practise explaining why a local candidate becomes global rather than merely naming the sign pattern.
Build the matrix before calculating any determinant. Place pure squared coefficients on the diagonal. Split each cross-term coefficient equally across the two symmetric off-diagonal positions, because both off-diagonal products contribute to the quadratic form. Expand x transpose A x once to verify that the original expression is recovered. Preserve the variable order and confirm symmetry.
This construction check prevents a doubled cross term from corrupting every later minor. Compute leading principal minors exactly and record their signs. Positive definiteness requires all of them to be positive. Negative definiteness requires alternating signs beginning negative. In two variables, that means a negative first leading minor and a positive determinant.
If a strict pattern fails, do not immediately declare the opposite classification. A zero minor can indicate a semidefinite case, while a negative determinant in a symmetric two by two matrix establishes indefiniteness. Use direct directions of the quadratic form as a check when the sign changes. Link the matrix result to the correct scale of optimisation claim.
A negative definite Hessian at a stationary point supports a strict local maximum. A Hessian that is negative semidefinite throughout a convex domain establishes concavity, while negative definiteness throughout supports strict concavity. Those domain-wide properties can turn valid first-order or Kuhn-Tucker conditions into a global conclusion, and strict concavity supports uniqueness.
State whether the evidence was checked only at one point or across the feasible domain. Use mixed revision sets that start from both directions. Given a quadratic expression, construct and classify its symmetric matrix. Given a Hessian, classify curvature and state the optimisation consequence that is justified. Include a semidefinite example so that a failed strict test does not trigger an automatic indefinite label.
Finish applied problems by checking that the feasible set is convex and by connecting any multiplier to the units of the binding resource. The final sentence should say maximum or minimum, local or global and unique or not established, matching the actual evidence. Build a curvature evidence table while practising.
Record the symmetric matrix, leading minors, definiteness class, domain over which the Hessian result holds and the resulting optimisation claim. Include one row where the strict test fails and additional analysis is required. Compare a pointwise negative definite Hessian with an everywhere negative definite Hessian on a convex domain.
The table makes clear why the former supports a strict local conclusion while the latter can establish strict concavity and a unique global optimum under the remaining conditions. Finish by expanding one quadratic form from its matrix and differentiating one objective to recover its Hessian.
These reverse checks connect coefficients, cross terms and curvature instead of treating definiteness as an isolated sign-pattern exercise.
Working through Definiteness and Concave Programming in ECON10071? Sia is AskSia’s AI Mathematics tutor — ask any ECON10071 Definiteness and Concave Programming question and get a clear, step-by-step explanation grounded in how ECON10071 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.