If I understand your question correctly, then:
a) a population is a set which interest you. the population in your case is all the houses in the UK. you are interested in the size of them. the parameter is an attribute of the population, usually it's something you do not know and want to find out. in this case, the parameter is the "average size of a UK household", which I would have defined "expectation" rather than "average", but you won't understand why at that stage. another definition of population can be the distribution of the data, but it's a more abstract definition.
b) in order to be 100% sure, and I repeat, 100% sure, you must check every house in your population, not other way.
c) since checking the entire population is not reasonable, some other procedure was probably taken. a sample, which is a subset of the population was taken, let's say, for example 300 houses. then an average of this sample was calculated, and was used for inference on the parameter we want to find out. inference means taking the sample, calculating some value from it ( called "Statistic", in this case the average, and using the statistic to predict the true value of the parameter we want, in this case the population's average.