Research Method for Managers

Thibault

LEMONNIER

BFS5

[pic 1]

Research Methods

 for Managers

Final Assignment

1. During the seminar, we worked with the following cases:

1. Alternic, Nikelen and Progesic products

In this case I learned to use the SUMPROD function, found variables, objective function and constraints and use the solver.

1. Computers S and P

For this exercise, I learned to use sensitive analysis to understand the limit and the shadow value of the case.

1. Hamlet: Make or buy

For this exercise, we had to helm Hamlet Inc. to decide how many products it should produce or purchase. Following different constraints on machines and costs we determine some plan for Hamlet Inc. In this exercise, I learned to combine two entries to have the best result.

1. Scheduling personnel

In this case, we had to help a Call Centre reducing costs of its workers. Considering schedules, we determine a planning where employees work 5 days a week. I learned to minimize the result for the Call Centre, choosing an option in the solver. To minimize costs, you must respect some constraints about time of work and the different wages between the days in the week and days in the week-end.

1. Water distribution network

This multi parametric exercise have a lot of variables, and the topic of this, is to maximize the flow in source nodes. In this case I learned to target my request to solve a problem. It was difficult due to the amount of information and constraints, different types of nodes, capacity, costs, demands.

1. Bus routes assignment

In this case, we had to assign a bus to a road. In this exercise, I learned to use a binary model to solve a question. The problem of this, was to understand how to convert the model to have the result in a binary version.

1. Sensitive analysis permit to optimise our objective function understanding how to modify as best as we can our constraints and our objective coefficient. You can move your value as much as the constraint allowed you, and your value move as much as the shadow price is. So, you can found the best result for you while respecting constraints.

[pic 2][pic 3]

1. The binary system give to solution only 1 if it’s true or 0 if it’s false. On many example, which use the binary systems there is a lot of 0 due to one solution if available for one given situation. Like the example with the bus, one road is assigned to one company, so on road 1 there is one company and one company can have at least 2 roads. That the reason why there is a lot of 0 on results. When we use a binary system, we can have a problem with value and you must put enormous values to show that solution isn’t possible due to the size of this value.

1. We would try to increase the constraint B because the range for allowable increase and allowable decrease is bigger than in the constraint A and you can move easily your constraint to adapt your result to the mood.
2. The shadow price of C is 0 due to the allowed increase is infinite and have no impact on the objective function. The shadow price on a non-binding constraint is zero.   If we have not used all of a resource available to us, then small changes in the right-hand side do not affect the optimal solution.
3. The maximal increase you can have on the constraint A is 1,67, so you can’t multiply by two. Even if you do, 1,67*0,67 stays under the constraint B.

1. Variables:

A1: amount of money invests on bonus A during the 1ST quarter

A2: amount of money invests on bonus A during the 2ND quarter

A3: amount of money invests on bonus A during the 3RD quarter

A4: amount of money invests on bonus A during the 4TH quarter

B1: amount of money invests on bonus B during the 1ST quarter