{"title":"Chapter 1: Introduction to Nonlinear Programming","authors":"","doi":"10.1137/1.9781611976199.ch1","DOIUrl":"https://doi.org/10.1137/1.9781611976199.ch1","url":null,"abstract":"A nonlinear program (NLP) is similar to a linear program in that it is composed of an objective function, general constraints, and variable bounds. The difference is that a nonlinear program includes at least one nonlinear function, which could be the objective function, or some or all of the constraints. Many real systems are inherently nonlinear, e.g. modelling the drop in signal power with distance from a transmitting antenna, so it is important that optimization algorithms be able to handle them. The problem is that nonlinear models are inherently much more difficult to optimize. There are twelve main reasons for this, as described below. Reason 1: It's hard to distinguish a local optimum from a global optimum. Numerical methods for solving nonlinear programs have limited information about the problem, typically only information about the current point (and stored information about past points that have been visited). The usual available information is (i) the point x itself, (ii) the value of the objective function at x, (iii) the values of the constraint functions at x, (iv) the gradient at x (the derivative in a 1-variable problem), and (v) the Hessian matrix (i.e. the second derivative in a 1-variable problem). This is enough information to recognize when you are at a local maximum or minimum, but there is no way of knowing whether there exists a different and better local maximum, or even how to proceed towards it. This also means that there is no way to easily determine where the global optimum is. These ideas are illustrated for a 1-variable unconstrained problem in Fig. 16.1. Formally, a local optimum x* is a feasible point that has a better value than any other feasible point in a small neighbourhood around it, say within some small ball with radius ε. The global optimum, on the other hand, is the point with the best value of the objective function anywhere in the feasible region. Note that the global optimum will also be a local optimum. Reason 2: Optima are not restricted to extreme points. There are a limited number of places to look for the optimum in a linear program: we need only check the extreme points, or corner points, of the feasible region polytope. This isn't so for nonlinear programs: an optimum (local or global) could be anywhere: at an extreme point, along an edge of the feasible region, or in the interior …","PeriodicalId":145755,"journal":{"name":"Practical Methods for Optimal Control Using Nonlinear Programming, Third Edition","volume":"124 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128145661","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Back Matter","authors":"","doi":"10.1137/1.9781611976199.bm","DOIUrl":"https://doi.org/10.1137/1.9781611976199.bm","url":null,"abstract":"","PeriodicalId":145755,"journal":{"name":"Practical Methods for Optimal Control Using Nonlinear Programming, Third Edition","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132316487","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Chapter 6: Partial Differential Equations","authors":"","doi":"10.1137/1.9781611976199.ch6","DOIUrl":"https://doi.org/10.1137/1.9781611976199.ch6","url":null,"abstract":"","PeriodicalId":145755,"journal":{"name":"Practical Methods for Optimal Control Using Nonlinear Programming, Third Edition","volume":"75 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134553005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}