Results of running Nordhaus original code using other solvers:
Knitro-IP: Got a solution which is good before 45 decades (But the first run failed: FUNC DOMAIN: x**c, x < 0 in equation YY(38))
Knitro-3: Failed in getting solution: FUNC DOMAIN: x**c, x < 0 in equation YY(2)
SNOPT: got a solution with a bit oscillation of emission control rate for the last periods, and capital solution becomes bad after 41 decades.
IPOPT: solution is good, almost the same with CONOPT
MINOS: got a solution in which emission control rate becomes bad after 47 decades, and capital solution becomes bad after 41 decades.