[Solved] GeneticAlgorithms Assignment2 Marketing Budget Allocation Problem

Marketing can help scale and sustain a business, but it comes with a cost! It is expensive to run paid campaigns, so careful planning is necessary. Given a marketing budget and a set of marketing channels, we would like to allocate the budget to the different marketing channels such that the total profit is maximized.

What you are required to do:

Write a genetic algorithm to solve the marketing budget allocation problem.

What the input looks like:

Youll be given the marketing budget, the set of marketing channels, the return-oninvestment (ROI) of each channel and a set of constraints (on investing in each channel). For simplicity, we will assume we have only 2 types of investment constraints:

• An investment in marketing channel (i) cant be less than some value (in thousands)
• And/or an investment in marketing channel (i) cant exceed some percent of the total budget.

These constraints form the lower and upper bounds of investment in each channel.

Example:

Notes:

• The ROI is given as a percentage.
• In this example, the investment in TV advertising cant be less than 2.7k and cant be more than 58% of the total budget. The investment in marketing through Google cant be less than 20.5k and there is no restriction on the upper bound. The investment in marketing through Twitter cant exceed 18% of the total budget and there is no restriction on the lower bound. The investment in marketing through Facebook cant be less than 10k and there is no restriction on the upper bound.
• Assume a single currency is used in the entire problem.
• Your final solution doesnt have to match the one in this example, but it should produce high profit. You could also display the output in a different format.

Important remarks:

1. The first thing you need to do is to think about how you will encode the solution in your chromosome. Formulate the problem with the objective function and constraints to help you figure out the next steps.
2. You should use a floating-point representation of the
3. The population size and initialization method you use are up to you. You can actually try different population sizes to see how this will affect your results. The maximum number of generations is also up to you.
4. Think about how you will handle infeasible solutions. Infeasible solutions are solutions that violate the constraints of the problem; therefore, they are not allowed.
5. Implement tournament selection, 2-point crossover and elitist replacement. Implement both uniform and non-uniform mutation.
6. Run the algorithm 20 times with the same input, write the output to a file and print the best solution obtained from these experiments.