Carryfast:
Firstly it’s the slow merging operation ‘at’ the obstruction that’s ‘causing’ the ‘queue’.While it’s obvious that the government are all about artificially slowing down the road network in every way possible.
Which is why we don’t see the lane closure signs put at least 1 mile or more before the obstruction and the road works sign at 800 yds so everyone can merge into the correct lanes at higher speeds well before the obstruction.As opposed to vice versa which means that the merging operation has to be done at a crawl. 
On that note,as it stands and without that change in the lane closure warning regime,it’s obvious that everyone just has to accept the situation of an artificially created bottle neck on the basis of slow zip merging at the obstruction rather than a high speed lane change to merge long before it.
“Nonsense”, I said to myself as I read this post, before noticing it was you Carryfast.
The only way a bottleneck could be avoided when two lanes (equally trafficked) merge into one, is if density is doubled, speed is doubled, or some combination of the two.
Speed will rarely increase, so usually density increases to a certain point, and then speeds drop to allow further density increases (as stopping distances are consumed), and finally (if necessary) traffic at the entry stops to queue at maximum density.
Obviously, merging modestly earlier (tens or hundreds of yards rather than miles) than the actual bottleneck is prudent so that traffic proceeds smoothly through the restriction at its optimum density and speed, rather than two jockeying cars hesitating or slamming on to avoid collision, which will then cause a gap to appear in the stream of traffic entering the restriction.
If traffic is free-flowing at motorway speeds, you should merge across at normal ‘sliproad’ distances of hundreds of yards. If traffic within the restriction is stationary, and a queue has formed behind it, you should proceed to within tens of yards and zipper merge.
The effect of merging much earlier than at the actual bottleneck, once a queue has already formed, is just to create exactly the same situation, but at an imaginary bottleneck further back.
It does not smooth the flow into the bottleneck, beyond the tens of yards already stated - and in fact in real life where a mixture of strategies are asserted, it may interrupt the flow, as two lanes approach at different speeds and then jockey about what is ‘fair’, whereas this cannot happen if the lanes are used equally to begin with (i.e. if my strategy is asserted dominantly, yours is neutralised, whereas yours when asserted dominantly not only fails to prevent the assertion of mine, but actually enables a new strategy: of queue-jumping along long, empty lanes).
What your strategy does do however is projects the queue backwards at higher speed, possibly requiring drivers to brake more harshly, and queuing will also reach previous junctions much sooner (i.e. literally causing gridlock).
The advice of traffic engineers stands that unless advised otherwise by signs, then all normally available road space should be used to queue (i.e. people should merge as required by encountering the bottleneck itself, not as required by encountering the queue formed in other lanes behind it), and traffic from each lane should then proceed in turn.
…(I aint doing myself any favours on this thread am I