Reverse Engineer Bank Balances - Commissions, Payment Dues and Errors - Algoritm Design

I need someone to develop a Learning Algoritm app. The problem is as follows.

Our Dealer (a big company) sells our product. Being big, they can alter the commission and payment dues on their behalf. We are fond of the partnership but they dont seem to care much on giving information on payments dues. There are two different systems ( DB: Our system, REAL: their system, hence the bank balance)

Here is how the system works:

1- a sale is made. An error here is less likely. Amount comes from here.

2-End of day is estimated. End of day is cut off whenever their cronjob works. It might be 20:00 or 03:00 next day. Thus, actual sales made on 3rd might be added to EndOfDay for 2nd or 4th. [Error 1] Last or first transactions of a specific day are effected from this error.

3-DueDate is estimated. Duedate is added to EndofDay. For instance, if transactions EndOfDay is 3rd and payment due is 10 days the due date is 13th. Two types of errors appear here. They forget (might be due to company holiday mismatch as well) to send a payment, but this is sent next day. [Error 2] They might change our payment due without asking to us. This means we dont get a payment for few days or get nearly double payment for few days. [Error 3]

4-Commission is estimated From every transaction, before they send the payment, and they first deduct their comission. They might change the comission rate without asking to us as well. [Error 4] This might be for a brief period or for an ongoing indefinite time.

5-Unknown returns: If a sale is returned by the consumer, they are supposed to inform us, so that we change our records (DB). However, sometimes they might not. But they deduct the associated amount from out payment. If the item is returned before duedate, then the amount is deducted at the duedate [Error 5]. If the item is returned after duedate, then the amount is deducted from the first payment [Error 6]

Every working day we see a transfer balance at the bank. We dont get to know which sales go into that payment.

From the series of the bank balances, we expect your Algoritm to derive,

REAL:

-Which sales go into which payment

-Comission

-Payment Due

And with these i expect the software to specify:

Which transactions encountered which error.

To simulate the environment you can assume:

-there are 50-200 transactions per day.

-amounts in the range of 2 to 2000 usd

-commissions range from 0 to 5 percent

-paymentdue might change from 1 to 60 days

I have included an excel file for a simple dummy data. But you are welcome to create on your own.

Steps for the project.

1- Create a sample transactions DB

2- Estimate expected bank balances from the db (DB Balance)

3- Create an algo for creating randomized errors for [Error1-6]

4- Estimate Real Balances (REAL Balance) considering the Errors in step 3

5- Without taking Errors into consideration, SOFTWARE should get REAL Balances and reverse engineer your way back to detailing Errors.

You can use any language or any platform. But JAVA and mysql is preferred.

The deliverables for the platform should include a working app which does the following:

1-Define a date period.

2-Clear and create a new transaction DB (with dummy data) within the date period

3-Add errors 1-6 with different variables. Add multiple of different errors.

4-Show DB Balance and REAL Balance for the period

5-Get REAL Balances and SOLVE for ERRORs (without using any data from item #3)

I have added a more complex file.
Normally u edit light blue cells, yellow and green are estimated.

But since we only know the bank balance (green cells)
THUS:
From
INPUT:the series of the bank balances (GREEN CELLS IN THE EXCEL),
we expect your Algoritm to derive,
OUTPUT: regular commission, regular PaymentDue, errors and parameters of errors. (YELLOW AND LIGHT BLUE CELLS IN THE EXCEL)

Pemrograman C Java PHP Perancangan Perangkat Lunak

ID Proyek: #8184930

Tentang proyek