Академический Документы
Профессиональный Документы
Культура Документы
SUBMITTED BY
TO:
1
Solver is an add in in Excel that helps in solving optimal solution in excel for
all types of decision problems. In a solver, the emphasis is on creating a
program or library that can easily be applied to other problems of similar type.
1.
2.
2
3.
3
4.
5.
4
Solver will be enabled.
5
This dataset required that find optimal profit by altering the number of calls
whole row in such a way that profit is maximised.
Here we maximize profit and thus max option is selected. Had this been a cost
or an error, we would have used min function as these would be mostly low.
Variable cells are defined as those values which could easily be altered or
varied in a way that the optimal objective solution is met, i.e. in this example,
number of calls are varied so as to get maximum profit. This value could be
singular or be in a range.
6
Constraints are put so as to either limit the variable cells so that they either
do not go beyond the predefined value, be equal to predetermined value, greater
than the predetermined value, that variable has to be always an integer or binary
number or that all variable cells must have different values.
1.
2.
3.
7
This means that reference value should be greater than or equal to 0
4.
5.
This means that cell reference must take only binary value
8
6.
GRG stands for “Generalized reduced Gradient”. In its basic form, this
solver technique looks on the gradient of the goal because the input
9
values (or decision variables) change and determines whether it has
reached a gold standard solution whilst the partial derivatives equal to
zero. Of the 2 nonlinear fixing methods, GRG Nonlinear is the fastest.
The downside is that the solution purchased with this algorithm is
enormously dependent on the initial conditions and won't be the ideal go
to solution. The solver will most probably prevent algorithm computation
once it reaches its optimal output when computation is nearest to the
initial conditions, providing with an answer that may or might not be
optimized globally.
In the earlier example, we used GRG Non linear and got optimal profit.
10
2. Simplex LP-
For example,
In this example, we observe that steel, labour and machine are linearly
dependent on resources and certain limit constraints are given. After
Randomly entering values, we use solver to optimize values to get the
optimised resource utilization
11
3. GRG Multistart-
Hybrid between the speed of the GRG Nonlinear algorithm and the robustness
of the Evolutionary algorithm is GRG Nonlinear Multistart. It can be enabled
this option through the Solver Options window, under the GRG Nonlinear tab.
The algorithm creates a randomly distributed population set of initial values that
are each evaluated using the traditional GRG Nonlinear algorithm.
4. Evolutionary Solver-
12
The Evolutionary algorithm is more robust than GRG Nonlinear due to
the fact it's far more likely to discover a globally best solution. however,
this solver technique is also VERY slow.
The Evolutionary technique is primarily based on the concept of herbal
selection – which works nicely in this situation due to the fact the most
fulfilling outcome has been described beforehand.
In simple phrases, the solver begins with a random “populace” of sets of
input values. Input values are plugged into the model and the effects are
evaluated relative to the target value.
The units of enter values that bring about a solution that’s closest to the
goal value are used to create a 2d population of “offspring”. The
offspring is a “mutation” of that excellent set of enter values from the
primary population.
This goes on until there may be little or no exchange within the goal
function from one population to the following.
What makes this method so time-consuming is that each member of the
population must be evaluated. Next “generations” are populated
randomly rather than the use of derivatives and the slope of the goal
function to locate the following excellent set of values.
13
We may choose the Mutation fee and populace length to probably shorten
the solution but, this has diminishing returns due to the fact reducing the
populace size and/or growing the mutation rate might also require even
greater populations to attain convergence.
14
We can keep the solver solution, and see detailed reports of answers,
sensitivity and limits also.
In the Convergence box, enter the amount of relative change that you
want to allow in the last ten iterations before Solver stops with the
message “Solver converged to the current solution.” Lesser values
usually mean that Solver might take more time, but will stop at a point
closer to the optimal solution.
15
used to specify the size of population and seed value is used for
specification of random seed number generator.
16