CoIoTe Project: optimization algorithm.

General description:

A number of different tasks:

• People are able to control the status of the dumpsters through a sensor system located into the dumpsters.
• People can collect data directly from the dumpsters and then they can inform the trash company and ask to empty the full dumpsters.
• People need to complete different tasks at different time steps during the day in a certain number of places and cities. The trash company can give people a prize/small payment in order to involve them to go in that zone.
• There are three different types of people: type 0, type 1 and type 2.

Each type can complete a different number of tasks in a time step and the cost of the type depends on how many tasks they can complete.

---

Goal:

The goal of the project is to minimise the total amount of the reward (rewarding people is a cost for the dumpster company).

---

Constrains:

Every dumpster must be visited; The total amount of requests cannot exceed the number of people in the cell.

---

Input file:

In the input file we have:

1. number of cells in which the city is divided;
2. number of time steps available;
3. tasks that must be solved in every cell at the end of all the time steps;
4. users available in each time step.

---

Variables:

There are nT different time steps
A city is divided in nC cells.
In some of the N cells there are a nA number of tasks to be solved.
In some of the N cells there also nP people that can complete those tasks.
People are divided in three type:

1. type 0 can complete 1 tasks every time step
2. type 1 can complete 2 tasks every time step
3. type 2 can complete 3 tasks every time step

---

Result:

On 120 instances we have an average of 0,15% from the optimal solution.

---

You can also find the PDF file with a more detailed explanation and others uploaded files.