| Author | Message | ||
     Lim Kok Han Melvin (lim_melvin) New member Username: lim_melvin Post Number: 1 Registered: 08-2003 |
I keep on getting the following error message, "Exponential Overflow". Is there anyway to avoid it? | ||
| Robert G. Nelson (rgnelson) Member Username: rgnelson Post Number: 11 Registered: 06-2003 |
The binary format in which numbers are encoded in the computer hardware places some limitations on the values that can be stored. In particular, the largest value that can be represented is approximately 10^308. This means that the largest acceptable value of the exponent in an exponential function is 709. The exponential function cannot return a value for an argument greater than this, and returns a hardware trap instead. Somewhere in your formulation, you have an exponential function whose argument is taking on very large values. Since it is unlikely that any number in the range of 10^308 is physically meaningful, I assume you have polynomial fits in the exponent, and these polynomials go insane outside of a narrow range of argument. The way to avoid this trouble is to use definitions which remain bounded within meaningful ranges, no matter what value the trial solutions may take. |