foi ueteimining a way to achieve the best outcome (such as maximum piofit oi lowest cost) in a given mathematical mouel foi some list of iequiiements iepiesenteu as lineai equations. Nany applications in mathematics involve systems of inequalitiesequations. Consiuei the following example: '~t_=/ -~, / = xample A fuinituie uealei ueals in only two items-tables anu chaiis. Be has Rs , to invest anu has stoiage space of at most pieces. A table costs Rs anu a chaii Rs . Be estimates that fiom the sale of one table, he can make a piofit of Rs anu that fiom the sale of one chaii a piofit of Rs . Be wants to know how many tables anu chaiis he shoulu buy fiom the available money so as to maximise his total piofit, assuming that he can sell all the items which he buys. Such type of pioblems which seek to maximise (oi, minimise) piofit (oi, cost) foim a geneial class of pioblems calleu optimisation pioblems. Thus, an optimisation pioblem may involve finuing maximum piofit, minimum cost, oi minimum use of iesouices etc. A special but a veiy impoitant class of optimisation pioblems is lineai piogiamming pioblem. The above stateu optimisation pioblem is an example of lineai piogiamming. Lineai piogiamming pioblems aie of much inteiest because of theii wiue applicability in inuustiy, commeice, management science etc. '~t_=/ -~, / n this example, we obseive that . The uealei can invest his money in buying tables oi chaiis oi combination theieof. Fuithei he woulu eain uiffeient piofits by following uiffeient investment stiategies. . Theie aie ceitain oveiiiuing conuitions oi constiaints viz., his investment is limiteu to a maximum of Rs , anu so is his stoiage space which is foi a maximum of pieces. Suppose he ueciues to buy tables only anu no chaiis, so he can buy - , i.e., tables. Bis piofit in this case will be Rs ( ), i.e., Rs .Suppose he chooses to buy chaiis only anu no tables. With his capital of Rs ,, he can buy - , i.e. chaiis. But he can stoie only pieces. Theiefoie, he is foiceu to buy only chaiis which will give him a total piofit of Rs ( ), i.e., Rs . Theie aie many othei possibilities, foi instance, he may choose to buy tables anu chaiis, as he can stoie only pieces. Total piofit in this case woulu be Rs ( + ), i.e., Rs anu so on. We, thus, finu that the uealei can invest his money in uiffeient ways anu he woulu eain uiffeient piofits by following uiffeient investment stiategies. Now the pioblem is : Bow shoulu he invest his money in oiuei to get maximum piofit. To answei this question, let us tiy to foimulate the pioblem mathematically. '~t_=/ -~, / + Let x be the numbei of tables anu y be the numbei of chaiis that the uealei buys. 0bviously, x anu y must be non-negative, i.e., x > y > The uealei is constiaineu by the maximum amount he can invest (Beie it is Rs ,) anu by the maximum numbei of items he can stoie (Beie it is ). Stateu mathematically, x + y < (investment constiaint)ie:- x + y < anu x + y < The uealei wants to invest in such a way so as to maximise his piofit, say, Z which stateu as a function of x anu y is given by Z = x + y (calleu objective function) Nathematically, the given pioblems now ieuuces to: Naximise Z = x + y subject to the constiaints: x + y < x + y < x > , y > So, we have to maximise the lineai function Z subject to ceitain conuitions ueteimineu by a set of lineai inequalities with vaiiables as non-negative '~t_=/ -~, / : } (Non-negative Constiaints) ..() ..() ..() ..() ..() '~t_=/ -~, / Let us giaph the constiaints stateu as lineai inequalities. The giaph of this system (shaueu iegion) consists of the points common to all half planes ueteimineu by the inequalities () to (). ach point in this iegion iepiesents a feasible choice open to the uealei foi investing in tables anu chaiis. The iegion, theiefoie, is calleu the feasible iegion foi the pioblem. veiy point of this iegion is calleu a feasible solution to the pioblem. The iegion othei than feasible iegion is calleu an infeasible iegion. \timal {feasible] solution: Any point in the feasible iegion that gives the optimal value (maximum oi minimum) of the objective function is calleu an optimal solution.Since theie aie infinitely many points, it is not eviuent how we shoulu go about finuing a point that gives a maximum value of the objective function Z = x + y. To hanule this situation, we use the following theoiems Theoiems eorem : Let R be the feasible iegion (convex polygon) foi a lineai piogiamming pioblem anu let Z = a+ Tbe the objective function. When Z has an optimal value (maximum oi minimum), wheie the vaiiables anu aie subject to constiaints uesciibeu by lineai inequalities, this optimal value must occui at a coinei point* (veitex) of the feasible iegion. eorem : Let R be the feasible iegion foi a lineai piogiamming pioblem, anu let Z = S+ Tbe the objective function. f R is bounded, then the objective function Z has both a maximum anu a minimum value on R anu each of these occuis at a coinei point (veitex) of R. '~t_=/ -~, / Coinei Point Nethou n the above example, the coinei points (veitices) of the bounueu (feasible) iegion aie: 0, A, B anu C anu it is easy to finu theii cooiuinates as (, ), (, ), (, ) anu (, ) iespectively. Let us now compute the values of Z at these points. We have We obseive that the maximum piofit to the uealei iesults fiom the investment stiategy (, ), i.e. buying tables anu chaiis. This methou of solving lineai piogiamming pioblem is iefeiieu as Coinei Point Nethou. '~t_=/ -~, / ' Applications of Lineai Piogiamming` .Manufacturing \roblems: n these pioblems, we ueteimine the numbei of units of uiffeient piouucts which shoulu be piouuceu anu solu by a fiim when each piouuct iequiies a fixeu manpowei, machine houis, laboui houi pei unit of piouuct, waiehouse space pei unit of the output etc., in oiuei to make maximum piofit. .Diet \roblems: n these pioblems, we ueteimine the amount of uiffeient kinus of constituentsnutiients which shoulu be incluueu in a uiet so as to minimise the cost of the uesiieu uiet such that it contains a ceitain minimum amount of each constituentnutiients. .rans\ortation\roblems: n these pioblems, we ueteimine a tianspoitation scheuule in oiuei to finu the cheapest way of tianspoiting a piouuct fiom plantsfactoiies situateu at uiffeient locations to uiffeient maikets. '~t_=/ -~, / llustiations ] {Diet \roblem] A uietician has to uevelop a special uiet using two foous P anu Q. ach packet (containing g) of foou P contains units of calcium, units of iion, units of cholesteiol anu units of vitamin A. ach packet of the same quantity of foou Q contains units of calcium, units of iion, units of cholesteiol anu units of vitamin A. The uiet iequiies atleast units of calcium, atleast units of iion anu at most units of cholesteiol. Bow many packets of each foou shoulu be useu to minimise the amount of vitamin A in the uiet. What is the minimum amount of vitamin A. olution: Let x anu y be the numbei of packets of foou P anu Q iespectively.0bviously x > , y > . Nathematical foimulation of the given pioblem is as follows: Ninimise Z = x + y (vitamin A) subject to the constiaints x + y > (constiaint on calcium), i.e. x + y > 8 ... () x + y > (constiaint on iion), i.e. x + y > ... () x + y < (constiaint on cholesteiol), i.e. x + y < ... () x > , y > ... () '~t_=/ -~, / / '~t_=/ -~, / // Let us giaph the inequalities () to ().The feasible iegion (shaueu) ueteimineu by the constiaints () to () is shown in the figuie. The cooiuinates of the coinei points L, N anu N aie (, ), (, ) anu (, ) iespectively. Let us evaluate Z at these points: Fiom the table, we finu that Z is minimum at the point (, ). Bence, the amount of vitamin A unuei the constiaints given in the pioblem will be minimum, if packets of foou P anu packets of foou Q aie useu in the special uiet. The minimum amount of vitamin A will be units. ] {rans\ortation \roblem] Theie aie two factoiies locateu one at place P anu the othei at place Q. Fiom these locations, a ceitain commouity is to be ueliveieu to each of the thiee uepots situateu at A, B anu C. The weekly iequiiements of the uepots aie iespectively , anu units of the commouity while the piouuction capacity of the factoiies at P anu Q aie iespectively 8 anu units. The cost of tianspoitation pei unit is given below: Bow many units shoulu be tianspoiteu fiom each factoiy to each uepot in oiuei that the tianspoitation cost is minimum. What will be the minimum tianspoitation cost. ///+=// -~, / /= ///+=// -~, / / olution:The pioblem can be explaineu uiagiammatically as follows: Let x units anu y units of the commouity be tianspoiteu fiom the factoiy at P to the uepots at A anu B iespectively. Then (8 - x - y) units will be tianspoiteu to uepot at C. Bence, we have x > , y > anu 8 - x - y > ie:- x > , y > anu x + y < 8 Now, the weekly iequiiement of the uepot at A is units of the commouity. Since x units aie tianspoiteu fiom the factoiy at P, the iemaining ( - x) units neeu to be tianspoiteu fiom the factoiy at Q. 0bviously, - x > , i.e. x < .Similaily, ( - y) anu - ( - x + - y) = x + y - units aie to be tianspoiteu fiom the factoiy at Q to the uepots at B anu C iespectively. Thus, - y > , x + y - > ie:- y< , x+y> Total tianspoitation cost Z is given by Z= x+ y+(-x) + ( - y) + (x + y - ) + (8 - x - y)= (x - y + 9).Theiefoie, the pioblem ieuuces to Ninimise Z = (x - y + 9) subject to the constiaints: x > , y > ... () x+y<8 ... () x< ... () y< ... () anu x+y> ... () ///+=// -~, / /+ '~t_=/ -~, / /: The shaueu iegion ABCBF iepiesenteu by the constiaints () to () is the feasible iegion.The cooiuinates of the coinei points of the feasible iegion aie (, ), (, ), (, ), (, ), (, ) anu (, ).Let us evaluate Z at these points. Fiom the table, we see that the minimum value of Z is at the point (, ).Bence, the optimal tianspoitation stiategy will be to uelivei , anu units fiom the factoiy at P anu , anu units fiom the factoiy at Q to the uepots at A, B anu C iespectively. Coiiesponuing to this stiategy, the tianspoitation cost woulu be minimum, i.e., Rs .