WRT the third one, this will probably flake out more people out as it's an obscure section of mathematics, but I'll try to put it simply.
In this example, we're talking about the number of lily pads at a point in time. Furthermore were modeling the situation such that time occurs in discrete intervals of number of days. At a particular point in time, or day, there are X lily pads and the time is measured in integer values. In other words, it's 1, 2, 3, etc not 1.5.
So at one point in time, let's call the days or time 'k" for simplicity we have X lily pads. At the next time, advancing one step in time or k+1 we have twice as many lily pads, or in other words we have 2X lily pads as we did at time k.
This lets us write the difference equation, y(k+1) = 2*y(k) where y is the number of lily pads (I changed to y from X because y is typically the dependent variable in algebra.
Now here's the trick. Going from time k to k+1 is an advance in time. Let's define this operation as an "advance" and call it the letter p. Therefore to advance in time, we operate on the function or multiply it by p. So now, p*y(k) = 2*y(k). In this particular system, there is no "forcing" function that impacts the number of lily pads and instead it's purely a function of the previous value. Therefore we only have what is called a complementary solution that can be found by solving the homogenous equation or p*y(k) - 2*y(k) = 0. Notice all we did is an algebraic subtraction. We can rewrites this by factoring out the y(k) as y(k)(p-2)=0. This equation had two solutions, either y(k) is zero or p=2. Either term would make the expression true, equaling zero). The solution of zero lily pads, y(k) =0 is called the trivial solution, therefore the solution we want is p=2. Notice here that that we've done is treat the time advance operation as an algebraic entity, which as Oliver Heavyside proved a long time ago, is a valid process.
Given the roots of the characteristic (homogenous equation) we know the solution takes the form of (root)^k or in this case 2^k or 2 raised to the k power. The actual generic solution is y(k) = A*2^k where A is a multiplicative constant determined by the initial conditions or in this case we know that at day 48 there is 100 percent coverage, let's call this 1.0. So now we can solve for A as saying y(48), which is 1 = A*2^48 or A = 1/(2^48). With this solution we can find the percent coverage at any day k.
There is a simpler way to look at it. Let's say that at day 0, the initial condition,there is 1 lily pad. Therefore 1 lily pad = A* 2^0. Since 2 raised to the zero power is 1 (anything to the zero power is identically one) 1=A*1 or A=1. Therefore the number of lily pads is y(k) =1*2^k.
Now we have the solution for all time, k or in table form
Time : lily pads
0 : 1
1 : 2
2 : 4
3 : 8
And so on. Notice that at each preceding time period we have half the number of lily pads. Therefore if at time k=48 we have 100% coverage we will half half the coverage one delay prior or at t=47.
The numbers work out for the other value of A and you will get .5 for k=47 or half coverage, but the constant is messy.
So there is you discrete time calculus lesson for today. This is what popped into my head on the way home yesterday.
Next time, we will solve this system with a forcing function that each day frog turds with seeds cause there to be an addition of an extra lily pad or y(k+1) = 2*y(k) + 1 which will have both a complementary solution and a particular solution with the general solution being the sum of the two.