Recent numerical integrations, however, suggest that stable planetary orbits exist: within three AUs (four AUs for retrograde orbits) of either Alpha Centauri A or B in the plane of the binary's orbit; only as far as 0.23 AU for 90 - degree inclined orbits; and beyond 70 AUs for planets circling both stars (Weigert and Holman, 1997).
We find that virtually all candidate systems are stable, as tested
by numerical integrations that assume a nominal mass - radius relationship.
Indeed, stable orbits may extend as far as one third of the closest separation between any two stars in a binary system, but according to NASA's Kepler Mission team,
numerical integration models have shown that there is a range of orbital radii between about 1/3 and 3.5 times the stellar separation for which stable orbits around two stars are not possible (Holman and Wiegert, 1999; Wiegert and Holman, 1997; and Donnison and Mikulskis, 1992).
These models vary in construction and contain different parameterisations of climate processes, and different methods for
the numerical integration (different grids, numerical schemes etc.).