@slypups
tl;dr: Whatever system you use, whether it's based on Wyoming or one of the systems used in countries that actually do proportional representation in their elections, there's going to be a big gap in people-per-representative between the "best" and "worst" states.
First off, if you're interested, here's an explanation of how House seats are shared out among the states: https://www.census.gov/population/apportionment/about/computing.html
Second, proportionality is a PRINCIPLE, and there are many, many methods to implementing it.
Setting Wyoming as the "unit" (or "quota" as election types technically call it) could work. Just divide a state's population by Wyoming's population (which was actually about 563k at the last census, not 523k) to get the number of quotas, and then if the quota has a decimal of 0.5 or above, round up to the nearest whole number, and if the quota has a decimal below 0.5, round down to the nearest whole number. Under this rule, California would have 66.08 quotas and get 66 seats. Texas would have 44.60 quotas and would get 45 seats. And so on.
But problem! The disproportionality at the margins would be almost as bad as what we have now. Under a Wyoming-quota system, South Dakota would have 1 representative for 841k people, and Delaware would get 2 reps for 898k people. That means SD would have 81% more people-per-rep than Delaware. Under the current system, Montana (with 989k people and 1 rep) has 88% more people-per-rep than Rhode Island (with 1.05M people and 2 reps). By comparison, 81% IS a mild improvement -- but emphasis on "mild."
[Side note: if the House had been re-sized in this way for the 2016 presidential election, the electoral college would have had 648 instead of 538, and Trump (assuming all electors voting faithfully) would have won 368-280.]
We could also try one of the methods that are used for proportional representation in other countries, while keeping the current House size at 435. There are four reasonably straightforward systems out there (many others also exist, including the current US one, but most others add complications to doing the calculation, usually with very minor differences in results, as you'll see):
1. The Hare quota. This isn't used much nowadays, but the US used it from 1850-1900. Its result: exactly identical to the current system.
2. The Droop quota, which is used in Ireland, Australia, and Malta. Its result: identical to the current system except that North Carolina would gain a seat, while Rhode Island would lose a seat. This would give Rhode Island 1 rep for 1.05M people, which is 70% more than Wyoming's 563k. This is a smaller gap than the both the Wyoming-quota rule and the current system would give you, but it's still a 70% disparity.
3. The Sainte-Laguë divisor, which the US used in 1840 and is used in several countries that have proportional representation today, such as Germany and New Zealand. Its result: identical to the Droop quota.
4. The d'Hondt divisor, which the US used from 1790-1830 and is currently used in most countries where proportional representation is practiced. Its result: quite a number of changes from the current system, with 8 of the 11 biggest states gaining seats and 7 of the 14 smallest states losing seats.....and Vermont and Wyoming getting NO SEATS AT ALL.