Software Packages

• SCALCG: Scaled Conjugate Gradient Algorithms for Unconstrained Optimization.

  testuo.pdf	Unconstrained Optimization Test Functions. Algebraic expression.
  evalfg.for	Fortran expression of test functions and their gradients.
  funcname.dat	Name of test functions.
  inipoint.for	Initial points specification.
  scalcg.for	SCALCG package - SCALed Conjugate Gradient Algorithm. Main program and subroutines.
  scg.for	SCG package - Birgin and Martinez. (Modified) Main program and subroutines.
  scwrev2.pdf	Scaled Conjugate Gradient Algorithm for Unconstrained Optimization (full paper).
  ss1.pdf	Computational results for SCALCG. theta spectral, stoptest=1, Powell restart.
  ss2.pdf	Computational results for SCALCG. theta spectral, stoptest=2, Powell restart.
  sa1.pdf	Computational results for SCALCG. theta anticipative, stoptest=1, Powell restart.
  sa2.pdf	Computational results for SCALCG. theta anticipative, stoptest=2, Powell restart.
  bm1sa.pdf	Computational results for SCG by Birgin & Martinez. theta spectral, stoptest=1, angle restart.
  bm2sa.pdf	Computational results for SCG by Birgin & Martinez. theta spectral, stoptest=2, angle restart.

  ---

• ASA: Anticipative Scalar Approximation of Hessian for Unconstrained Optimization.

  testuo.pdf	Unconstrained Optimization Test Functions.
  func.for	Fortran expression of test functions.
  asa.for	Fortran code with Anticipative and a variant of Barzilai-Borwein algorithms.
  paper.pdf	A new gradient descent method with an anticipative scalar approximation of Hessian for unconstrained optimization (full paper).
  asa.pdf	Numerical experiments with a gradient descent method using an anticipative scalar approximation of Hessian for unconstrained optimization.
  bb.pdf	Numerical experiments with a variant of Barzilai-Borwein method for unconstrained optimization.
  asabb.pdf	Performance profiles: Anticipative algorithm versus Barzilai-Borwein algorithm, 310 problems.

  ---

• RELAXED: Relaxed Gradient Descent Method with Backtracking for Unconstrained Optimization.

  testuo.pdf	Unconstrained Optimization Test Functions.
  func.for	Fortran expression of test functions.
  relax.pdf	Relaxed Gradient Descent Method with Backtracking for Unconstrained Optimization (full paper).
  relaxed.pdf	Numerical experiments with Relaxed Gradient Descent Algorithm. 430 problems.
  grad.pdf	Numerical experiments with Gradient Descent Algorithm. 360 problems.

  ---

• UNO:
  Fortran subroutines library for unconstrained optimization using searching methods. The following methods are implemented: Univariate probing technique
  Hooke-Jeeves
  Powell, Rosenbrock
  Nelder-Mead
  Chazan-Miranker in conjunction with unidimensional methods: Fibonacci
  Golden section
  Powell's quadratic interpolation
  Davidon's cubic intrerpolation.

---

• SPENBAR:
  Sparse PENalty BARrier package for solving general nonlinear programming problems with equality and inequality constraints and simple bounds on variables. All the constrained are penalized. Minimization of the penalty function is given by TN package of S. Nash.

---

• ALLO:
  A Language for Linear Optimization. This is a language for linear programming modeling. The associated translator is able to convert an algebraic description of a linear programming problem into the standard MPS form.
Algebraic Modeling Languages

---

• RF: Product Form of Inverse (PFI) with a preassigned pivot procedure.
  This is a package of subroutines dedicated to compute the PFI with a preassigned pivot procedure of the A**(-1), to solve the corresponding systems of linear equations, A*X=B or AT*X=B, exploiting sparsisty in all cases.

  Summary of available software for sparse direct methods

  rfpp.for	Fortran subroutines for PFI with a preassigned pivot procedure computation: -determine a permutation of rows and columns of the matrix to "bump-and-spike" structure, -compute the "etavectors" of the matrix permuted to "bump-and-spike" form, -froward transformation of RHS term, -backward transformation of RHS term, -insitu permuation of a vector, -compress the arrays with the matrix.
  sys.dat	Input file with some examples.

---