I have already programmed a football manager game some years ago with a lot functionality (Sent it to Gamestar for free publishing but was rejected because of copyrights).
I have one section in the game on that I am not really satisfied because I have no clue what would be the best fit to model it.
I have a vector of pairs for all days of one calendar year:
std::vector<std::pair<Competition*, TDate>> year_plan;
E.g. there can be following entries inside pointing to a Competion class instance and a date:
CHAMPIONS_LEAGUE_DRAW, 12.03.2022
CHAMPIONS_LEAGUE, 15.03.2022
FIRST_LEAGUE, 19.03.2022
SECOND_LEAGUE, 19.03.2022
--> As you can see, there can be several competitions at the same day!
The program logic is just processing the year_plan date by date and takes the action required on the specific competition on simulating results or simulating the draws. This perfectly works, but somehow it does not feel right this way.
For that reason I have two questions on that:
Would it not make more sense to have a one year Calendar class and instances for every Competion occurence in it? (One instance per day is not enough because of having soemtimes more than one competitions per day). This would also allow to store additional information in the instance like counter for the matchday, round etc.
The Competitions I currently have are LeagueCompetions, CupCompetitions, Draws. All these three competitions types have some parts in common but need also special class members and methods. Draw competitions need much less information to be stored. As I need to store the Competition in the Calendar vector / class, I need somehow one class forr all three competition types. Should I use here inheritance approach even I need a specialization of the classes in addition to the common parts?
Thanks in advance for any suggestions
Mauro
Related
I try to distribute x jobs among y persons using reinforcement learning (DQN).
Every person can have a specific amount of tasks and every task can only be done once.
I mask out all the non possible task for each person for example if a task is already choosen it will just be masked out (So the output size stays the same)
I preprocess my data by combining the features of the person with the features of the task. For example I would substract the timeslots: A Person has 4 timeslots left and the task needs 2 the resulting feature would be 2. I do this for every person and with every task resulting in one big matrix where the #rows = #persons and #colums = #tasks * #features.
Now I want to give my network as many information as possible meaning the whole matrix but I am unsure on how to do it.
One possible idea would be to make one big flatten array but the problems would be that the amount of persons can change and also that I can only choose one task at a time for one person so I would need to tell the network which person is the active one.
Another approach would be something like "Hey I have a sequence lets use RNN" but I am not sure how to teach the network which is the current person. I also think this would lead the network to give me the best task over all persons. But it should learn something like "If the task is better for another person don't choose it for the active one".
The output of my network are the actions(tasks) where I choose the maximum.
Maybe some smart person has an idea. Thanks for your help.
At my work we have two systems, one that collects the customers payments automatically every month. And one that manages the memberships of those customers. Sadly our outdated technology doesn’t communicate to each other so we don’t know if a customer actually paid for their membership without manually auditing them.
I’ve been put in charge of this process and boy does it take awhile to do.
I have limited knowledge of C++ and was looking into maybe writing a program to do the comparisons for me.
I have two ideas on how to implement this, and was wondering what you guys thought. If these would be best or if it’s even possible or if there’s a better solution?
Current Setup: We have a list of all members in excel, with how much each should be paying, we then go through the actual money collected and check to make sure everyone’s payment went through and was processed and not declined.
Option 1: have a multi-dimensional array of strings. Read the excel file into this array it would have three Columns, first name, last name, amount they should be paying. This would be put in alphabetical order to help with the searching. I would then export the transactions in css file format and read each line one at a time. When it reads a line it would search the array for the same first and last name. Once found it would take the amount paid confirm it said processed and not declined and if so would subtract it from the customers amount they should be paying. In the end if every customers amount they should be paying is equal to 0 then everyone paid.
Option 2: is similar to option 1 just instead of using a multidimensional array it would use two css files. And not put the items into the array at the start.
Thoughts? Is this a smart way to combat this problem? I’m a newbie programmer so I’m just looking for suggestions/advice.
Your solutions would work, but are suited for small datasets. I don't now what your constraints are, but I think that a more elegant solution would be to setup a database on the first system first(instead of the excel file).
Are you allowed to create a database? How many customers are in the excel file?
I have asked a few questions related to this personal project of mine already on this platform, and this should be the last one since I am so close to finishing. Below is the link to a mock example spreadsheet I've created, which mimics what my actual project does but it contains less sensitive information and is also smaller in size.
Mock Spreadsheet
Basic rundown of the spreadsheet:
Pulls data from a master schedule which is controlled/edited by another party into the Master Schedule tab.
In the columns adjacent to the imported data, an array formula expands the master schedule by classroom in case some of the time slots designate multiple rooms. Additional formulas adjust the date, start time, and end time to be capped within the current day's 24-hour period. The start time of each class is also made to be an hour earlier.
In the Room Schedule tab, an hourly calendar is created based on the room number in the first column, and only corresponds to the current day.
I have tested the spreadsheet extensively with multiple scenarios, and I'm happy with how everything works except for the calculation time. I figured the two volatile functions I use would take some processing time just by themselves, and I certainly didn't expect this to be lightning-fast especially without using a script, but the project that I am actually implementing this method for is much larger and takes a very long time to update. The purpose of this spreadsheet is to allow users to find an open room and "reserve" it by clicking the checkbox next to it (which will consequently color the entire row red) allowing everyone else to know that it is now taken.
I'd like to know if there is any way to optimize / speed up my spreadsheet, or to not update it every time a checkbox is clicked and instead update it "manually", similar to what OP is asking here. I am not familiar with Apps Script nor am I well-versed in writing code overall, but I am willing to learn - I just need a push in the right direction since I am going into this blind. I know the number of formulas in the Room Schedule tab is probably working against me yet I am so close to what I wanted the final product to be, so any help or insight is greatly appreciated!
Feel free to ask any questions if I didn't explain this well enough.
to speed up things you should avoid usage of the same formulae per each row and make use of arrayformulas. for example:
=IF(AND(TEXT(K3,"m/d")<>$A$1,(M3-L3)<0),K3+1,K3+0)
=ARRAYFORMULA(IF(K3:K<>"",
IF((TEXT(K3:K, "m/d")<>$A$1)*((M3:M-L3:L)<0), K3:K+1, K3:K+0), ))
=IF(AND(TEXT(K3,"m/d")=$A$1,(M3-L3)<0),TIMEVALUE("11:59:59 PM"),M3+0)
=ARRAYFORMULA(IF(K3:K<>"",
IF((TEXT(K3,"m/d")=$A$1)*((M3-L3)<0), TIMEVALUE("11:59:59 PM"), M3:M+0), ))
I'm using the AWS Machine Learning regression to predict the waiting time in a line of a restaurant, in a specific weekday/time.
Today I have around 800k data.
Example Data:
restaurantID (rowID)weekDay (categorical)time (categorical)tablePeople (numeric)waitingTime (numeric - target)1 sun 21:29 2 23
2 fri 20:13 4 43
...
I have two questions:
1)
Should I use time as Categorical or Numeric?
It's better to split into two fields: minutes and seconds?
2)
I would like in the same model to get the predictions for all my restaurants.
Example:
I expected to send the rowID identifier and it returns different predictions, based on each restaurant data (ignoring others data).
I tried, but it's returning the same prediction for any rowID. Why?
Should I have a model for each restaurant?
There are several problems with the way you set-up your model
1) Time in the form you have it should never be categorical. Your model treats times 12:29 and 12:30 as two completely independent attributes. So it will never use facts it learn about 12:29 to predict what's going to happen at 12:30. In your case you either should set time to be numeric. Not sure if amazon ML can convert it for you automatically. If not just multiply hour by 60 and add minutes to it. Another interesting thing to do is to bucketize your time, by selecting which half hour or wider interval. You do it by dividing (h*60+m) by some number depending how many buckets you want. So to try 120 to get 2 hr intervals. Generally the more data you have the smaller intervals you can have. The key is to have a lot of samples in each bucket.
2) You should really think about removing restaurantID from your input data. Having it there will cause the model to over-fit on it. So it will not be able to make predictions about restaurant with id:5 based on the facts it learn from restaurants with id:3 or id:9. Having restaurant id there might be okay if you have a lot of data about each restaurant and you don't care about extrapolating your predictions to the restaurants that are not in the training set.
3) You never send restaurantID to predict data about it. The way it usually works you need to pick what are you trying to predict. In your case probably 'waitingTime' is most useful attribute. So you need to send weekDay, time and number of people and the model will output waiting time.
You should think what is relevant for the prediction to be accurate, and you should use your domain expertise to define the features/attributes you need to have in your data.
For example, time of the day, is not just a number. From my limited understanding in restaurant, I would drop the minutes, and only focus on the hours.
I would certainly create a model for each restaurant, as the popularity of the restaurant or the type of food it is serving is having an impact on the wait time. With Amazon ML it is easy to create many models as you can build the model using the SDK, and even schedule retraining of the models using AWS Lambda (that mean automatically).
I'm not sure what the feature called tablePeople means, but a general recommendation is to have as many as possible relevant features, to get better prediction. For example, month or season is probably important as well.
In contrast with some answers to this post, I think resturantID helps and it actually gives valuable information. If you have a significant amount of data per each restaurant then you can train a model per each restaurant and get a good accuracy, but if you don't have enough data then resturantID is very informative.
1) Just imagine what if you had only two columns in your dataset: restaurantID and waitingTime. Then wouldn't you think the restaurantID from the testing data helps you to find a rough waiting time? In the simplest implementation, your waiting time per each restaurantID would be the average of waitingTime. So definitely restaurantID is a valuable information. Now that you have more features in your dataset, you need to check if restaurantID is as effective as the other features or not.
2) If you decide to keep restaurantID then you must use it as a categorical string. It should be a non-parametric feature in your dataset and maybe that's why you did not get a proper result.
On the issue with day and time I agree with other answers and considering that you are building your model for the restaurant, hourly time may give a more accurate result.
I have a booking system for something where the price can change based on the day. The admins for the site can make these changes. If a booking crosses the boundary of a daily rate, they pay pro-rata for the rates they used.
I'm losing confidence in how this is implemented. There are at least two ways:
Having Rates that specify their validity (start, end fields) and then working out which of those apply. But which overlapping ones take priority? Etc. Nasty. This is what we're trying to do and cannot currently answer sufficiently well.
The same except that there is some form of unique quality to date so that no two rates can overlap. The problem here is we'd need to split existing Rates on insert and rejoin two on delete/edit, etc if they had the same value. We'd need to make sure there were no gaps. It requires some heavy ORM overriding.
Keeping a DayRate table with every day defined. This means keeping a load of extra data around but most bookings are for tens of days, not thousands so I'm not worried about the database bandwidth requirements here. Date would be primary-unique and I'd just do a range filter for grabbing which ones I need to factor in.
The problem is generating these dates ahead of time. I know that as soon as I implement this, somebody will make a booking for 2032. Is there a good way around this or should we limit them?
None of these answers seems great and I have to imagine that I'm not the first guy with a booking system. Is there a better way of keeping track of a rate over a contiguous (possibly infinite) amount of time?