Академический Документы
Профессиональный Документы
Культура Документы
Optimisation is the act of obtaining the best result under given circumstances. In the design , construction and maintenance of any engineering system, engineers have to take many technological and managerial decisions at several stages. The goal of such decisions is either to minimise the effort required or to maximise the desired benefit.
Optimisation can be defined as the process of finding the conditions that give the maximum value or minimum value of a function. A number of optimisation methods have been developed for solving different types of optimsation problems
Historical Development The existence of optimsation methods can be traced to the days of Newton, Lagrange, Cauchy, Leibnitz, Bernoulii, Euler etc., Very little progress was made until the middle of the twentieth century. Then High speed digital computers made implementation of optimisation procedures possible.
Von Neumann gave the foundations for game theory in the year 1928 Development of Simplex method by Dantzig in 1947 Kuhn and Tucker conditions for NLP in 1951 Bellmanns principle of optimality in 1957
Developments in Numerical methods of unconstrained optimisation have been made in the UK in the 1960s.
Historical Developments (contd.,) Duffin, Zener and Peterson developed Geometric Programming during1960s. Gomory gave his cutting plane method for Integer programming. Goal programming was proposed by Charnes and Cooper in 1961. Simulated annealing, genetic algorithms and neural network methods came during the last decade.
3.
5. Design of water resources systems for maximum benefit 6. Optimal design of structures
7. Optimum design of linkages, cams, gears, machine tools and other mechanical components
11. Optimum design of electrical machinery such as motors, generators, and transformers 12. Optimum design of electrical networks
18. Planning of maintenance and replacement of equipment to reduce operating costs 19. Inventory control
20. Allocation of resources or services among several activities to maximise the benefit
Stochastic Techniques
Markov Processes
Nonlinear Programming
Geometric Programming Linear Programming Dynamic Programming
Queuing Theory
Renewal Theory Simulation methods Reliability Theory
Design of Experiments
Integer Programming
Stochastic Programming Separable Programming Multi objective Programming
Classification of Optimisation problems (contd.,) (iv) Based on the nature of equations involved:
Linear Programming
Nonlinear Programming
Geometric programming
Quadratic programming
Linear Programming
This is applicable for the problems in which the objective function and the constraints are linear functions of the key variables. Applications: Product mix problem Optimal production planning in a manufacturing firm Make or buy decisions in metal working industries Minimising trim loss in paper mill Optimal routing of messages in communication network Routing of aircraft and ship Transportation problem and assignment problem
Sensitivity
Analysis/Post
optimal
analysis
To know the sensitivity of the optimality of a LP model with respect to the changes made in the parameters and characteristics of a LP model
(i) (ii) (iii) (iv) Making changes in the RHS constants of the constraints Making changes in the objective function coefficients Adding a new constraint Adding a new variable -Also the resources used in the model may be classified as abundant and scarce and unit worth of a resource may also be calculated.
Nonlinear Programming
- It is an extension of linear programming. - The objective function may be nonlinear and the constraints may be linear or nonlinear.
Lagrangean method
Lagrangean method is used to solve a NLP where the objective function is nonlinear and the constraint is in linear equation form.
Kuhn-Tucker conditions
KT conditions are used to solve a NLP where the Objective function is in nonlinear form and constraint is in nonlinear inequality form.
Quadratic Programming
Quadratic Programming is a NLP problem in which the highest order of the polynomial of the objective function is restricted to 2.
Such quadratic programming problems are solved using Wolfes method. The QPP is converted into a LPP and solved using suitable simplex method.
Separable Programming
Separable programming problem is a NLP where the objective function consists of linear and nonlinear term as well as the constraint will also consist of linear and nonlinear term. In such problems the nonlinear terms are converted into equivalent sum of linear terms using piece wise linearisation and the problem becomes a LPP and it is solved using simplex with restricted basis method.
Integer Programming
Integer Linear programming is a LPP where the optimal solution for the key variables are to be obtained in terms of integers. Methods used: Gomoris cutting plane method Branch and Bound method
Dynamic programming
-Also known multistage programming -Bellmanns optimality condition is used. Terms used in DP: Stage State Variable Recursive function (Forward/Backward) More amount of computations to be made
Network Techniques
(i) CPM and PERT Project networks (i) Shortest path model
Sequencing
Optimal sequence of a few jobs to be processed under a few machines so that elapsed time and the idle times are minimised. Types: N Jobs x 2 machines N jobs x 3 machines N Jobs x m machines m jobs x 2 machines
Game Theory
-Two person Zero sum game -Maxi-min and Mini-max criteria -Pure strategy and Mixed strategy -Value of the game and the optimal strategies are found out. Types: (i) Games with saddle point (ii) Method of sub games (iii) Graphical method (iv) Matrix method
Queuing Models
Characteristics of a queuing model Arrival pattern Service pattern No. of service channel Service order No. of customers served Population
1
2
0.712
1.03
0.3516
0.415
4
4.5
2100
2300
Initial RPN
Sl.No.
1
Failure Mode
Low Tensile strength
RPN
160
Weld crack
120
Due to the FMEA that was carried out for friction welding in the manufacture of engine valves, the RPN for the different failure modes were reduced and are shown below. RPN Values
Improved RPN 32 40
Conclusion
- The optimisation techniques available are plenty. - A given problem must be properly understood to identify the function to be optimised, key variables to be optimised, constraints if any and accordingly the problem has to be formulated. - A suitable optimisation technique/model must be selected accordingly to derive the maximum benefit from these available techniques.
Thank you