GUIDELINES
for the Selection of Benchmarks
Two main areas of interest are suggested, each involving both homogeneous and non-homogeneous systems, namely:
 
1) Small nonlinear model problems (max dimension: 2-d, 3-d )
 
2) Large problems involving detailed chemical kinetics
 
Small nonlinear model problems
 
The candidate benchmark problem should feature:
 
1) tunable stiffness in the homogeneous case, to be relevant to the detailed chemistry problems;
 
2) internal layers representatives of reaction-diffusion coupling as in flames;
 
3) either linear  (e.g. Semenov) or nonlinear (e.g. chain branching) behavior at infinity.
 
Candidate benchmark problems can also feature some of the following characteristics:
 
a) a hyperbolic equilibrium point with (i) only real, (ii) only complex, or (iii) real and complex eigenvalues;
 
b) a non-hyperbolic equilibrium point (e.g. Lindemann);
 
c) a stable limit cycle (e.g.  Semenov);
 
d) a bifurcating cascade of limit cycles of  higher periods leading to chaotic behaviors (e.g. Lorenz or Franceschini-Tebaldi);
 
e) linear or nonlinear diffusion, as well as differential diffusion.
Problems with detailed chemical kinetics
The candidate benchmark problem should feature:
 
1) kinetic phenomena typical of complex hydrocarbon oxidation (e.g. two-stage ignition, negative temperature coefficient behavior);
 
2) large dimension, e.g. between 100 and 200;
 
3) chemkin-type specification format;
 
4) available on the net.
Candidate benchmark problems can also feature some of the following characteristics:
 
a) an equilibrium point;
 
b) a stable limit cycle.