Chapter 6 Linear Programming: The Simplex Method6.1 Geometric Introduction to the Simplex MethodGeorge Dantzig 1914 - 2005An Interview with George Dantzig, Inventor of the Simplex MethodAn Interview with George Dantzig (continued)Slide 6Slide 7Standard Maximization Problem in Standard FormExampleExample (continued)Slide 11Slack VariablesSlack Variables (continued)Example (continued)Slide 15Discovery!GeneralizationConclusionChapter 6Linear Programming: The Simplex MethodSection 1A Geometric Introduction to the Simplex Method26.1 Geometric Introduction to the Simplex MethodThe geometric method of linear programming from the previous section is limited in that it is only useful for problems involving two decision variables and cannot be used for applications involving three or more decision variables. It is for this reason that a more sophisticated method must be developed.3George Dantzig1914 - 2005George B. Dantzig developed such a method in 1947 while being assigned to the U.S. military. Ideally suited to computer use, the method is used routinely on applied problems involving hundreds and even thousands of variables and problem constraints.4An Interview with George Dantzig, Inventor of the Simplex MethodIRV How do you explain optimization to people who haven't heard of it? GEORGEI would illustrate the concept using simple examples such as the diet problem or the blending of crude oils to make high-octane gasoline. IRVWhat do you think has held optimization back from becoming more popular? GEORGEIt is a technical idea that needs to be demonstrated over and over again. We need to show that firms that use it make more money than those who don't. IRVCan you recall when optimization started to become used as a word in the field? GEORGEFrom the very beginning of linear programming in 1947, terms like maximizing, minimizing, extremizing, optimizing a linear form and optimizing a linear program were used. http://www.e-optimization.com/directory/trailblazers/dantzig/interview_opt.cfm5An Interview with George Dantzig (continued)GEORGEThe whole idea of objective function, which of course optimization applies, was not known prior to linear programming. In other words, the idea of optimizing something was something that nobody could do, because nobody tried to optimize. So while you are very happy with it and say it's a very familiar term, optimization just meant doing it better than somebody else. And the whole concept of getting the optimum solution just didn't exist. So my introducing the whole idea of optimization in the early days was novel. IRVI understand that while programming the war effort in World War II was done on a vast scale, the term optimization as a word was never used. What was used instead? GEORGEA program can be thought of as a set of blocks, or activities, of different shapes that can to be fitted together according to certain rules, or mass balance constraints. Usually these can be combined in many different ways, some more, some less desirable than other combinations. Before linear programming and the simplex method were invented, it was not possible to computationally determine the best combination such as finding the program that maximizes the number of sorties flown. Instead, all kinds of ground rules were invented deemed by those in charge to be desirable characteristics for a program to have.6An Interview with George Dantzig (continued)A typical example of a ground rule that might have been used was: "Go ask General Arnold which alternative he prefers." A big program might contain hundreds of such highly subjective rules. And I said to myself: "Well, we can't work with all these rules." Because what it meant was that you set up a plan. Then you have so many rules that you have to get some resolution of these rules and statements of what they were. To do this, you had to be running to the general, and to his assistants and asking them all kinds of questions. IRVName some of your most important early contributions. GEORGEThe first was the recognition that most practical planning problems could be reformulated mathematically as finding a solution to a system of linear inequalities. My second contribution was recognizing that the plethora of ground rules could be eliminated and replaced by a general objective function to be optimized. My third contribution was the invention of the simplex method of solution.7An Interview with George Dantzig (continued)IRVAnd these were great ideas that worked and still do. GEORGEYes, I was very lucky. IRVWhat would you say is the most invalid criticism of optimization? GEORGESaying: "It's a waste of time to optimize because one does not really know what are the exact values of the input data for the program." IRVOk, let's turn this around. What would you say is the greatest potential of optimization? GEORGEIt has the potential to change the world.8Standard Maximization Problem in Standard FormA linear programming problem is said to be a standard maximization problem in standard form if its mathematical model is of the following form:MaximizeP = c1x1 + c2x2 + … + cnxnsubject to problem constraints of the forma1x1 + a2x2 + … + anxn < b, b > 0with nonnegative constraints x1, x2, …, xn > 09Example We will use a modified form of a previous example. Consider the linear programming problem of maximizing z under the given constraints. This is a standard maximization problem.5 108 8 1604 12 1800; 0z x yx yx yx y= ++ �+ �� �10Example(continued) There are four lines bordering the feasible region:000124088yxyxyxAny two of them intersect in a point. There are a total of six such points.The coordinates of the point can be found by solving the system of two equations in two unknowns created by the equations of the two lines.Some of the points are corner points of the feasible region, and some are outside.11Example(continued)grid spacing = 5 units(0,0) feasible(0,15) feasible(0,20) not feasible(7.5,12.5) feasible (optimal)(20,0) feasible(45,0) not feasible12Slack Variables To use the simplex method, the constraint inequalities must be converted to equalities. Consider the two constraint inequalities 18012416088yxyxTo make this into a system of equations, we introduce slack variablesThey are called slack variables because they take up the slack between the left and right hand sides of the inequalities. Note that the slack variables must be