Population growth

Based on two data points, you can't. There is not anywhere near enough information to make a reasonable estimate.

the population will be higher in 2020, than it was in 1999 and 2000

barring an unforeseen event (major disease, nuclear war, meteorite strike)

Barrin unforseen events like illness, accident, and body parts wearing out, I will live to be 120.

You can find the difference between the two yearly figures and extrapolate that value to the rest of the years. It won't be very accurate but it's the best you can do with that data and no additional information.

LET THE POPULATION OF YEAR 1999 BE m AND THAT OF YEAR 2000 BE n , ASSUME 1999 TO BE 1(=t) IN A NEW TIME SCALE DENOTED AS "t" THEN FOR t=2 FOR YEAR 2000 AND t=22 FOR YEAR 2020. ASSUME THAT
THE POPULATION(y) AT ANY TIME t > 0 SATISFIES y= a( t^u) + b( t^v) WHERE a AND b ARE CONSTANTS TO BE DETERMINED AND u AND v ARE INTEGERS "AT ONE S OWN CHOICE". CONSTANTS a AND b ARE DETERMINED BY SOLVING m=a +b AND n= a (2^u) + b (2^v) SINCE,"m" AND " n" ARE KNOWN, AND WE GET THE POPULATION AT YEAR 2020 AS y=a (22^u) + b (22^v) . THE MAIN DRAWBACK OF THE SAID CONSIDERATION IS THE DETERMINATION OF THE INTEGERS u AND v WHICH ARE LEFT AT ONES OWN CHOICE WHICH IS NOT MATHEMATICALLY RIGOROUS,EVEN THE RELATION,AT FIRST ASSUMED BETWEEN y AND t MAY NOT HOLD.STILL I THINK THAT THE ABOVE METHOD PROVIDES AN APPROXIMATELY SATISFACTORY METHOD FOR DETERMINING THE POPULATION.