This preview shows pages 1–4. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: 1 Making Differential Equations &onDeterministic Dan Gillespie Dan T Gillespie Consulting GillespieDT@mailaps.org Current Support : Caltech (NIGMS, NIH) University of California at Santa Barbara (NIH) Past Support : Caltech (DARPA/AFOSR, Beckman/BNCM)) University of California at Santa Barbara (DOE) Molecular Sciences Institute (Sandia/DOE) Office of &aval Research Its everywhere ( ) ( ) ( ), dX t A X t t dt = . (1) Newtons 2 nd Law: 2 2 d x F m dt = ( ) ( ) ( ) ( ), ( ), ( ) dx t v t dt F x t v t t dv t dt m = = Defining 1 ( ) ( ) ( ) and ( ( ), ) ( ) ( ( ), ( ), ) v t x t X t A X t t v t m F x t v t t , we get (1). If the independent variable t represents time , we call X a process .  Time is a very special variable. It always increases . 2 the Past the Present the Future ( ) ( ) ( ), dX t A X t t dt = (1) Another way of writing (1), which shows explicitly that it determines the future values of X in terms of the present value of X , is this: ( ) ( ) ( ( ), ) ( ) ( 0) X t t X t A X t t t o t t + = + + , where ( ) o t satisfies lim ( ) t o t t + = . Or more simply . . . ( ) ( ) ( ( ), ) X t dt X t A X t t dt + = + . (2) In (2), dt is an ordinary real variable , distinct from t, which is confined to [0, ), where is an arbitrarily small positive number. h In other words . . . dt is a little bit of t . time ( ) ( ) ( ), dX t A X t t dt = (1) ( ) ( ) ( ( ), ) X t dt X t A X t t dt + = + . (2) In any real situation, (2) will be a sufficiently accurate expression of (1) for all [0, ] dt for some sufficiently small > . (1) and (2) are logically equivalent , because they define the same process X . (2) has the form of an update formula for X : Its a recipe for computing, from the value of X at the present time t , its value at the later time t dt + for all [0, ] dt . 3 ( ) ( ) ( ( ), ) X t dt X t A X t t dt + = + (2) The process X defined by (2) has three obvious properties : X is continuous : ( ) ( ) X t dt X t + as dt . X is memoryless (Markov) : To compute any future value ( ) X t dt + , (2) needs only ( ) X t ; it does not need any past values ( ) X t t < . X is deterministic : The update formula fixes the future value ( ) X t dt + unequivocally. A Question : Is (2) the only form for an update formula for a process that is continuous and memoryless and deterministic ? How about: ( ) ( ) ( ( ), ) X t dt X t A X t t dt + = + . (*) There is no problem taking the square root of any number in [0, ) . X as defined by (*) is not differentiable . But so what?...
View
Full
Document
This note was uploaded on 07/08/2009 for the course ME 210B taught by Professor Linda during the Spring '09 term at UCSB.
 Spring '09
 Linda

Click to edit the document details