The Student Room Group

Edexcel D1 - 17th June, 2016 [Exam discussion]

Scroll to see replies

Original post by gekkelekkebekke
same here mate, let''s hope they're not gonna be ***** on friday


Which was a very hard D1 paper that many found diff?
I'm an idiot and have only given myself tonight and tomorrow to learn from scratch the entire unit. Somebody said that exam technique was the most important thing with this exam, but I have no teacher and can't seem to find anything about it. Does anyone know anything about the specific ways they're expecting you to format answers, such as with the algorithms?
Original post by taysc
Forget about S1. It is possible to do crap in one unit and do amazing in another unit to balance out your grade. You still have a little bit of time to practice even more, go in with the confidence that you know the knowledge. I always think they can't go too far of the specification and if they do you have to use common sense or think back to the basics of the topic.

You will do great as long as you know your stuff. I always suggest trying to finish 20-30 mins early to check over your work and make sure you said things like 'sort complete'. Make sure you state what your doing, marks are still given for simple explanations e.g. state why you choose to focus on arcs excluding 'K' and checking floats on cascade charts/checking you have scheduled activities according to what they rely on.

Trust me speaking positivity can completely turn a situation around, hope your last bits of revision go well.


I did well last year apart from M1 haha. Thanks for the kind words of encouragement. Much appreciated. Still sucks I have D1 and M2 on the same day :biggrin:
D1 should not be hard for anyone doing A2. Hardly touched it this year but will do about 15 hours tomorrow, which should be ok. I think we should be worried for the C3 and C4, looking at what people have been saying about the C3 of AQA.
D1 NOTES OFF MEMORY:
Just seeing how well i can remember topics on this unit. Tomorrow is just a pure past paper day - I've got some videos to watch by UKmathsteacher then the PP grind is real. I have tried some and I'm getting around 50-60ish marks and want to improve that by 10.

Algorithms:
Bubble sort - simply look at each pairing in the list and undertake appropriate swaps - you move left to right and can only swap each pair once. It is nice to underline fixated / finalised items and state that the sort is complete as no more passes need to be made.

Quick sort - Looks like examiners love seeing how you define your pivots via using N+1/2 or N+2/2 depending if it is odd or even. Identify and state pivots. Again move left to right and place items the order they appear in surrounding these pivots. It is nice to circle and then put a square around pivots or some other method of showing it clearly. Think I've got it pretty smooth, although I could probably do some more practise to just improve my speed.

Full Bin = Mix and match yourself to find optimal solution of bins.
First fit bin packing = Place numbers in bins in the order they appear in.
First fit decreasing = Numbers in descending order then place in the order they appear in.
With regards to formatting this, usually drawing a rectangle to literally represent the bin/ stacking values in one another seems acceptable.

Binary search :
Show clearly all of the pivots you use and why you reject certain items. Remember the list must be in an order firstly and that you always delete the pivot itself. Also, the list essentially always halves.

Kruskal's = List arcs in ascending order with respect to their weight,
Prim's = Grow from a point. Tabular form I feel comfortable with due to it appearing in D2.

Chinese postman : My favourite one.
*Extra - Handshake theorem implies that each line added to a network adds 2 valency therefore a network must always have an even valency.*
You CANNOT be asked more than 4 nodes.
Identity odd nodes. Look at the different pairings available - 4 odd nodes have 3 different pairings. List the paths you undertake to obtain the smallest possible values and see which pairing gives the least value. The smallest weighted pairings are the ones that you want to traverse twice as they add the minimal amount of weight to the network. Remember if the smallest route from A-C is A-B-C then we would want ABC traversed not just AC!

Dijkstras : Blergh
First thing I always do is trial and error myself and get a grasp as to what kind of small value I am aiming for. Just be slow and methodical about the process - don't do what I sometimes do is and forget to write down working values. Put them all down and persevere.
Current - previous = weight I+J then it lies on the path. Is this correct? I can't recall the actual formula for it - memory is slacking!

Matchings :
Unpaired on X vertex to unpaired on Y vertex. Find the path and make sure to change state and clearly show the examiner this. The algorithm improves the solution by one each time. When asked for why is a matching not available, usually there is clear limitations with the tasks being assigned I.e 2 people may only be able to do one task therefore there can be no complete matching. I quite like these ones, as long as they are not unnecessarily long.

Linear Programming: good and bad
- When given loads of information and you have to formulate it as a linear programming problem, I always draw a table and it makes life easier.
- Plotting the lines assign either X or Y zero and find values.
- remember that usually X,y are larger than or equal to zero is a constraint.
- Just do a lot of point testing to be certain with inequalities - sometimes it is obvious, but better be safe to check.
- Test each vertex and use ruler method. You usually want whole integers. Sometimes you need to look at points of interesection between 2 lines to determine the exact values if needed.

It feels like a daunting topic because I've done so little practise on it but can be alright sometimes.

Critical path : Why why

Lower bound for workers = summation of arcs / last critical value

Drawing a network from a precedence table - literally just watch UKmathsteachers videos on it, that's what clarified it for me. Can be confusing and trial and error to get a nice looking graph - kind of defeats the idea of it being maths if none of the answers are unique. It did help me understand the idea of dependency though.

Critical path table values:
Right to left = smallest
Left to right = biggest
Be careful with dummy's the most.

Float = Li - ej - duration i,j and those with float of zero are critical.
Always check every point - list the floats if every point, it makes gantt charts more bearable.
I think that remembering that formula for float basically means you know 3/4 of critical path.
I do not like scheduling - it's more trial and error. Any tips ? I've heard just use guidelines to use those with the smallest time and just keep at it.
Gantt charts do not show dependency clearly.

That's a rough guide I hope. I've probably missed some things but I think that sometimes just writing everything out like this is a great way to see how much I know. It's a weird unit and I find it hard in some places- just hope I can bring my marks up well tomorrow. Best of luck all ! Please correct me if need be.
Reply 325
Original post by TrueDAN
...I do not like scheduling - it's more trial and error. Any tips ?...


In general, the most effective way is to assign tasks based on the value of their late event time (lowest value of late event time goes first). This video helped me to understand how to do it: https://youtu.be/46SGt7O8So8

I don't like these either... they take up way too much time for the marks allocated.
(edited 7 years ago)
Does any one have the january 2016 IAL paper for D1?
Original post by smackitintheair
Does any one have the january 2016 IAL paper for D1?


https://56leomessiphotoshop.blogspot.com/p/edexcel-january-2016-exam-materials.html


It would be 21<=x<=30

If you do that and put 31 as the value, you'll see you could have two solutions i.e. not unique any more, but even if it was 30, you would get the same solution not a different one
Anyone come across any really tough Critical Path Analysis & Scheduling questions? Would really appreciate.
Original post by hogree
Anyone come across any really tough Critical Path Analysis & Scheduling questions? Would really appreciate.


jan 2016 IAL
Original post by js.int
In general, you choose the most effective way is to assign tasks based on the value of their late event time (lowest value of late event time goes first). This video helped me to understand how to do it: https://youtu.be/46SGt7O8So8

I don't like these either... they take up way too much time for the marks allocated.

Thanks a lot! :smile:
It's never come up before, but I was just wondering how having say 2 workers to each activity would change the scheduling diagram?
Reply 336
Original post by N_ght.yo
It's never come up before, but I was just wondering how having say 2 workers to each activity would change the scheduling diagram?


Because there's no algorithm per se for the numbers of workers they could ask any question that uses basic analysis of the gantt diagram etc
Anyone pls define

Matching
Early Event time?
What does named correctly mean? Do you have define each letter you pick? Also when doing a quick sort can you do it like this? Sorry for the bad paint use haha
Is there a way of checking Dijkstra's algorithm?

Quick Reply

Latest

Trending

Trending