REMEMBER TO ENTER INTEGERS WITH TWO DECIMAL POINTS (eg.15.

00)
Module 1: Module 2:
Linear programming: find best way to achieve goal (maximize or minimize Sensitivity analysis: manipulate objective function/coefficient of isoprofit line
objective) before optimal solution changes.
Linear Statements: Variables without exponents or square root sings Change to RHS of Binding Constraints
Variables/Action Plan: Unknown but must be decided (each variable = decision) - If change is allowable increase/decrease, each unit of increase/decrease
Objective function/target cell: measure using quantities (profit/cost) will change target cell (total profit) by whatever value the price of the
Regular Constraints: A limitation of some sort posed with the problem. constraint is.
Input Values: data given The optimal solution will change
Quantitative Decision - Using data and other relevant information, often in some - If there is an additional cost…
type of model, in order to find answers to business (and other) problems. - Allowable decrease * shadow price = x
Objective function – the function describing the problem’s objective which you are - Cost of decreasing value = y
attempting to maximize or minimize. - X - y = target cell change
Objective Coefficient - Normally the prices of objective functions - Any change outside the allowable range will lead to a change to both
Optimal solution – the best set of decisions that maximizes the objective function optimal solution & target cell (we don't know exactly how these will
while remaining within the constraints. change)
How many optimal solutions with LP? 0, exactly 1, infinity Change to RHS of non-binding Constraints
Constraint – A limitation of some sort posed with the problem. - Any increase or decrease to the RHS that is within the allowable range
Multiple optima – There are multiple sets of optimal solutions. will not change the target cell nor the optimal solution.
Non-negativity constraint – A constraint that makes sure a “decision” cannot be a - Any change outside the allowable range will lead to a change in both the
negative value. solution and target cell, but we don't know how exactly they will change.
Binding non-negativity: anything zero within red cells Within allowable increase: Shadow price * units + target value
Binding Constraint: LHS=RHS (all resources used) (graphically: only binding if Final value * (increase or decrease) x shadow price
they intercept) Change Objective coefficient within its allowable range: optimal solution (red
Non-Binding Constraint: LHS ≠ RHS (something is unused) (shadow price=0) cells) doesn’t change & target cell (green cell) does. Range in which the objective
(allowable increase: infinity (1E+30) when LHS < RHS ) (allowable decrease: coefficient may move without disrupting the optimal solution.
RHS-LHS) the line doesn’t touch optimal solution Shadow Price – The increase in the value of the target cell for every one-unit
Redundant constraint – A constraint which does not affect the feasible region. increase of the RHS of a constraint.
Revenue= price * quantity Shadow Price equation: new profit - old profit
Feasible region – The region in which all solutions are valid and subject to the EXAMPLE: if oil A output increased by 10L how much will profit change?
constraints. Shadow price=0.1 Increase= 10 L ANSWER= 0.1 x 10(shadow price x Increase)=
Infeasible solution – There is no feasible region associated with your LP. No area 1$
that satisfies all constraints. No optimal solution. Sensitivity report in excel: When your model is finished, run the solver, and then
Unbounded solution – The feasible region is infinitely large, usually due to lack of answer this question: What settings did you use in the solver dialog box? Make
a constraint, and the objective function behaves such that you are moving the unconstrained variables non-negative AND the Simplex LP solving method.
isoprofit line outwards indefinitely. Sometimes no optimal solution: max(no cause it Sensitivity report(find change in target cell):
keeps going up to infinity) Min(yes cause u can go back to original point) (final value)(new objective coefficient) - (final value)(objective coefficient)
Relative Reference – A reference in the form A1 that will change when auto-filled Isoprofit Line- Graphically represents the objective function. Identifies which
to other cells. corners of feasible region is optimal solution (better way then)
Absolute Reference – A reference in the form $A$1 that will not change when Slope: -C1/C2 (use objective function)
auto-filled to other cells. Slope is negative: move isoprofit line to right Slope is positive: move to left
Assumptions typically made in simple models- There is unlimited demand... we Main property: the value of the objective function stays the same along the
can sell as many products as we make. isoprofit line
You can tell what type of graph if you see: NON-BINDING CONSTRAINTS DONT HAVE A SHADOW PRICE
Scheduling(shifts), Blending(%) Transportation(supply, demand, Allowable increase of a on binding constraint is 1e+30 (infinty)
received, shipped) Simple(lego model)-what to produce/eat/ Changing a binding constraint within its allowable range usually has the following
no complications effect: Optimal solution and target cell both change.
In transportation models supply does not have to equal exam.
Changing an objective coefficient within its allowable range usually has the
following effect: Optimal solution doesn't change and target cell changes.

How many optimal solutions is it possible to have with an LP? 0,1 and infinity.