Re: Short Quadratic Solver (HP-42S) Message #6 Posted by Csaba Tizedes (Hungary) on 24 Oct 2010, 6:08 p.m., in response to message #5 by Gerson W. Barbosa
Maybe the HP's SOLVE algorithm is accurate?
Here is my solution (from my memories...) - not shortest, uses variables, but not uses the classical quadratic formula. The 32SII (low-end) version:
--- The firs root:
I01 LBL I
I02 CF 0
I03 FN=Q
I04 SOLVE X
I05 GTO X
I06 SF 0
I07 RTN
7 steps / CK=368D / 10.5 byte
--- The second:
X01 LBL X
X02 STOP
X03 RCL B
X04 RCL div A
X05 +
X06 +/-
X07 RTN
7 steps / CK=0339 / 10.5 byte
--- The quadratic equation:
Q01 LBL Q
Q02 RCL X
Q03 RCL mul A
Q04 RCL add B
Q05 RCL mul X
Q06 RCL add C
Q07 RTN
7 steps / CK=8AAE / 10.5 byte
'add', 'sub', 'mul' and 'div' means 'add', 'substact', 'multiple' and 'divide' in this order.
Solve for real roots of an 'a*x^2+b*x+c=0' equation:
a STO A, b STO B, c STO C then XEQ I.
If FLAG 0 set this equation has no real roots (location of extremum on display).
If FLAG 0 not set, the first real root appear on LCD, then press R/S, then the second root will be on the display.
Enjoy!
|