12 replies to this topic

[xxpineapplexx]

Does anyone know how to seed a casio fx9860GII so when i do randint my answers are the same as my professors?

[flyingfisch]

I don't understand what you mean. Your professors will get random numbers and so will you. How can you get the same results as your professors when you are generating random numbers?

[Forty-Two]

http://en.wikipedia....iki/Random_seed

pinapple: a quick search through the the manual shows nothing on seeding. You could try implementing your own and calling that whenever you need a random number, but I don't see any other way... :l

[xxpineapplexx]

In class my professor uses TI 83 calculators and he can seed the calculators so they can spit out random numbers from a list. The seed is the line that the calculator spits out the numbers from. Its like a random number table in the calculator because as you know the a machine can never give truly random numbers. Because all the calculators in class are seeded to the same line, the random numbers every student gets are the same.

[flyingfisch]

You could probably make a simple BASIC prog to do that...

[MicroPro]

Of course this is difficult enough. This page brought by google search suggests the calculator uses L'Ecuyer's algorithm and google has things about it. If you figure out how it works (any TI dudes have?) you may be able to implement this in Basic.

Meanwhile manually copy the random numbers from your friend's TI to make sure you have the same random numbers.

[xxpineapplexx]

Thanks for all the suggestions so far and I don't want to sound ungrateful or anything but I think you guys are overthinking this. In class the students with TI's use the following commands to seed their calculators: They put the line number ,like 5, then they input STORE, and finally they input Randint and then the calculator is seeded. I assume that there should be a similar command for my far superior Casio.

[Forty-Two]

Thanks for all the suggestions so far and I don't want to sound ungrateful or anything but I think you guys are overthinking this. In class the students with TI's use the following commands to seed their calculators: They put the line number ,like 5, then they input STORE, and finally they input Randint and then the calculator is seeded. I assume that there should be a similar command for my far superior Casio.

Yes, because this functionality is supported in the TI-83. It is not in the fx. Please understand this.

[flyingfisch]

So there is a list and the calc randomly chooses a number from the list?

You could make a BASIC prog to do that like this:

Locate 1,1,List 1[RandInt(1,Dim List1)]

You could put the numbers into list one and then that would return one.

It could all be done from the homescreen, too.

You would do this (linear mode):

{1,2,3,4,5} -> List 1 // (numbers to be seeded)
Prog "SEED" //name of seeding program

That would give you the number right in the graph screen.

[Forty-Two]

So there is a list and the calc randomly chooses a number from the list?

You could make a BASIC prog to do that like this:

Locate 1,1,List 1[RandInt(1,Dim List1)]

You could put the numbers into list one and then that would return one.

It could all be done from the homescreen, too.

You would do this (linear mode):

{1,2,3,4,5} -> List 1 // (numbers to be seeded)
Prog "SEED" //name of seeding program

That would give you the number right in the graph screen.

Nooooo...

The idea is that there is a function that takes a seed and outputs a number. You start with a non-random seed, and feed that in toget the first 'random' number. For a given seed, the same number will always be outputted . To get the next number in the series, take the previous random number and use it as a seed. Note that the output of the rng is usually larger than the output to the user. After a while, the rng will end up with the first number being produced. The amount of time it takes to do this is called the Period of the rng.

To demonstrate, here's a very simple rng with a period of 4 in pseudocode.

static nibble seed= 0x3 //0011

bit random(){

seed = seed << 1;  //This shifts the seed over 1 bit, wrapping
return seed >>> 3;  //This shifts the seed over 3 bits, nonwrapping.  This essentially selects the leftmost bit. 

}

Lets run the following code:

println(random());
println(random());

Initially, seed=0101. On the first line of random(), seed gets set to 0110. The return value is the leftmost bit, or 0. The next time random() is run, seed gets changed to 1100. The leftmost bit this time is 1. There are a couple problems with this rng, including a very low period, high dependance of starting value of the seed, and low length of output, but it is good for an example.

[flyingfisch]

ah, i see. so you would have to know what kind of pseudo random number generator TI calcs use, right?

[Forty-Two]

As MicroPro said, it uses L'Ecuyer's algorithm.

[flyingfisch]

ah, ok. So where can we find it? this search doesn't seem to have much...