5.

MODELE DINAMICE
5.1. Metoda controlului optimal
)HQRPHQHOH H[SOLFLW QXPLW L SURFHVHOH HFRQRPLFH DX ORF vQ WLPS /XDUHD vQ FRQVLGHUDUH VSHFLDO

D WLPSXOXL FRQGXFH OD PRGHOH GLQDPLFH FDUH VH UH]ROY

FX R PHWRG

FRQWURO RSWLPDO 0HWRGD FRQWUROXOXL RSWLPDO D IRVW FRQFHSXW

SHQWUX WUDWDUHD

PRGHOHORU GLQDPLFH FRQWLQXH (D D IRVW SXV

OD SXQFW vQ DQLL FLQ]HFL SHQWUX DSOLFD LL

PLOLWDUH L LQGXVWULDOH GH F WUH PDWHPDWLFLDQXO UXV 3RQWUHDJKLQ 8WLOL]DUHD PHWRGHL FRQWUROXOXL RSWLPDO vQ HFRQRPLH SHUPLWH R DERUGDUH WLLQ LILF D SUREOHPHORU

LQWHUWHPSRUDOH L FRQWULEXLH OD FODULILFDUHD PRGXOXL GH DUWLFXODUH D FRQVLGHUD LLORU SH WHUPHQ OXQJ L VFXUW

1. &DUDFWHUL]DUH JHQHUDO
0RGHOXO GLQDPLF

D PRGHOHORU GLQDPLFH HVWH R UHIOHFWDUH D VLVWHPHORU HFRQRPLFH GLQDPLFH L VH

caUDFWHUL]HD]

SULQ XUP WRDUHOH HOHPHQWH

D 2UL]RQWXO L VFDUD GH WLPS 6LVWHPXO HFRQRPLF HVWH VWXGLDW L DQDOL]DW SH R DQXPLW SHULRDG GH WLPS

T.

Definim, deci, orizontul de timp (sau de planificare) ca intervalul [0,T]. Un decupaj al acestui interval n timp.
Q ILJXUD GH PDL MRV HVWH GDW R DVWIHO GH VFDU  SHULRDGH HOHPHQWDUH GH R DQXPLW OXQJLPH GHWHUPLQ VFDUD GH

--|------|------|------|-------------|------|------------|-------> 0 h 2h 3h ........ (k-1)h kh ...... nh

1XP UXO WRWDO GH SHULRDGH HOHPHQWDUH

n, este definit de: n=T/h, unde T este

lungimea orizontului, iar h este lungimea perioadei elementare. Cea de a k-D SHULRDG
HOHPHQWDU vQFHSH OD PRPHQWXO

(k-1)h L VIkUHWH OD PRPHQWXO kh.

Pentru variaELOHOH
GH QRWD LL

PRGHOHORU GLQDPLFH VXQW XWLOL]DWH vQ PRG X]XDO GRX

WLSXUL

QRWD LD LQGLFLDW  FDUH UHSHUHD]

YDULDELOHOH PRGHOXOXL SULQ LQGLFHOH SHULRDGHL YDULDELOHOH vQ IXQF LH GH WLPSXO VFXUV GH OD

QRWD LD IXQF LRQDO  FDUH UHSHUHD]

debutul orizontului de timp.

7UHFHUHD GH OD XQ WLS GH QRWD LH OD DOWXO SHUPLWH WUDWDUHD ULJXURDV GLQWUH WLPSXO GLVFUHW L WLPSXO FRQWLQXX D UDSRUWXOXL

b) Variabile de stare
8Q PRGHO GLQDPLF XUP UHWH HYROX LD vQ WLPS D IHQRPHQXOXL DFHDVW HFRQRPLF PRGHODW 'HRVHELP GRX WLSXUL GH YDULDELOH FH FDUDFWHUL]HD] FH VH vQWkPSO HYROX LH

YDULDELOH WLS SHULRDG YDULDELOH

IOX[  FDUH P VRDU

FX IHQRPHQXO PRGHODW vQWU

-o
DO

HOHPHQWDU  WLS VWRF  FDUH P VRDU HIHFWXO FXPXODW OD XQ PRPHQW GDW

HYROX LHL SURFHVXOXL

VariabileOH
REVHUYDW

WLS

VWRF

H[SULP

starea

IHQRPHQXOXL

PRGHODW $FHDVW

VWDUH HVWH

OD PRPHQWHOH

{0, h, 2h,...,kh,...,nh=T}. La data t=kh, starea procesului este -un vector din R
p
SUHVXSXQHP F H[LVW

UHSUH]DQWDW

SULQ YDULDELOH JUXSDWH vQWU

variabile care caUDFWHUL]HD]

VWDUHD OD XQ PRPHQW GDW 1RW P DFHVW YHFWRU GH VWDUH  VDX SULQ

prin X(t) vQ QRWD LD IXQF LRQDO

p

Xk vQ QRWD LD LQGLFLDW

 R DQXPLW HYROX LH

3H RUL]RQWXO DOHV IHQRPHQXO VDX SURFHVXO PRGHODW XUPHD] vQ VSD LXO

R al variabilelor de stare, reSHUDW

SULQ LUXO GH YHFWRUL

{X(0),X(h),...,X(t-h),X(t),...,X(T)},
sau

{X0,X1,...,Xk-1,Xk,...,XT/h}.
(YROX LD SURFHVXOXL PRGHODW vQ VSD LXO VW ULORU SRDUW QXPHOH GH

traiectorie.

c) Variabile de control (sau de decizie) n timpul perioadei (t-h,t] asupra unor variabile tip "flux" trebuie luate decizii,
DGLF WUHEXLH DOHV QLYHOXO DFHVWRUD 0XO LPHD YDULDELOHORU GH WLS IOX[  QXPLWH L FRUHVSXQ] WRDUH XQHL SHULRDGH HOHPHQWDUH GDWH

variabile de control sau de decizie, sunt regrupate ntr-un vector din R

1RW P DFHVW YHFWRU SULQ XW VDX GH FRQWURO OD GDWD W SRDWH IL OLPLWDW

SUHVXSXQHP F

VXQW

m variabile de decizie).

uk $OHJHUHD GH F

WUH GHFLGHQW D QLYHOXOXL YDULDELOHORU

GH XQ QXP U GH UHVWULF LL 6FULHP DFHVW OXFUX VXE

forma u(t)U sau ukU

0XO LPHD

HVWH R SDUWH D VSD LXOXL

QXPLW

PXO LPH GH V GHSLQG

GHFL]LL DGPLVLELOH Q XQHOH PRGHOH GLQDPLFH HVWH SRVLELO FD PXO LPHD GH WLPS L GH VWDUHD VLVWHPXOXL

U = U(X(t),t).

G (FXD LD GH WUDQ]L LH Q HYROX LD VD SURFHVXO WUHFH GH OD R VWDUH OD DOW GHSLQGH GH VWDUHD vQUHJLVWUDW GDW OD GDWD

a. Starea sistemului la data t

GHSHQGHQ VH VFULH vQ IRUP

t-h, de nivelul variabilelor de control la acea

u(t) L GH OXQJLPHD SHULRDGHL HOHPHQWDUH h $FHDVW

DVWIHO

JHQHUDO

X(t)=A(X(t-h),u(t),t,h) , t=h,2h,...,T
p

[1]

unde A () HVWH R DSOLFD LH IXQF LH FX YDORUL vQ R  )RORVLQG QRWD LD LQGLFLDW

 DYHP

Xk = Ak ( Xk-1, uk, h ) , k=1,2,...,T/h

5HOD LD >@ GHILQHWH HFXD LD GH WUDQ]L LH FDUH DUDW PRGXO FXP VLVWHPXO WUHFH

dintr-o stare n alta.

2 FODV IRDUWH U VSkQGLW GH PRGHOH GLQDPLFH VXQW FHOH GLIHUHQ LDOH DGLF FHOH vQ FDUH HFXD LD GH WUDQ]L LH HVWH R HFXD LH FX GLIHUHQ H ILQLWH GH WLSXO

X(t)-X(t-1)= a(X(t-h), u(t), t)h , t=h,2h,...,T

[2]

sau Xk - Xk-1 = ak (Xk-1, uk)h , h=1,2,...,T/h.

)XQF LD

ak ()

P VRDU

YDULD LD VW ULL VLVWHPXOXL vQ XQLWDWHD GH WLPS D SHULRDGHL

k 'HFLGHQWXO FRQWUROHD]

HYROX LD VLVWHPXOXL DF LRQkQG vQ ILHFDUH SHULRDG  QX GLUHFW

DVXSUD VW ULL FL LQGLUHFW DVXSUD YDULD LHL

acesteia.

H 6WDUHD LQL LDO

D VLVWHPXOXL HVWH FXQRVFXW 

La data t=0 starea sistemului, X0

VHFYHQ D GH GHFL]LL

X0=.

'DF

VH FXQRDWH L

{u1,u2,...,uk,...,uT/h} atunci traiectoria sistemului este perfect

L DPG Q XQHOH FD]XUL HVWH FRQVWUXLP

GHWHUPLQDW

SULQ UH]ROYDUHD HFXD LLORU GH WUDQ]L LH >@ VDX >@ SDV FX SDV FXQRVFkQG

X0

u1 se deduce X1, cunoscnd X1

FXQRDWHP VWDUHD ILQDO

u2 se deduce X2

SRVLELO V

XT=, L vQ ED]D VHFYHQ HL GH GHFL]LL V

DVHP Q WRU WUDLHFWRULD Q SULPXO FD] VH ]LFH F FHO ODOW PRGHOXO HVWH UHWURVSHFWLY 'HVLJXU F HVWH GLIHULW  VWDUHD VLVWHPXOXL OD GDWD

PRGHOXO GLQDPLF HVWH SURVSHFWLY vQ

vQ FD]XO UHWURVSHFWLY HFXD LD GH WUDQ]L LH

t depinznd de starea sa la data t+1.

f) Criteriul de optimizare
'HFLGHQWXO XUP UHWH XQ RELHFWLY GHWHUPLQDW 'H H[HPSOX HO SRDWH DYHD GUHSW VFRS V PD[LPL]H]H XQ FULWHULX GH IRUPD

J=[

Bk(Xk-1,uk)]h + M(XT/h)

sau J=[B(X(0),u(1),1)+..+B(X(t),u(t+h),t+h)]h+..+M(X(T)).
)XQF LD DD FXP HVWH L P VRDU OD GDWD FkWLJXO UHDOL]  $FHVWD OD

Bk(Xk-1,uk)
HO GH VWDUHD

at, n unitatea de timp, n perioada k,

GH GHFL]LD OXDW vQ SHULRDGD DFHOHL SHULRDGH )XQF LD

HYDOXDW

GHSLQGH

UHVSHFWLY P VRDU

VLVWHPXOXL

vQFHSXWXO

M(XT/h)
R VXP

YDORDUHD DWULEXLW

VLVWHPXOXL HFRQRPLF OD VIkUL

tul orizontului de timp.

HVWH

n majoritatea modelelor economice dinamice, criteriul J

DFWXDOL]DW YDORULORU DUDW F SHQWUX R SHULRDG GH OXQJLPH

GH IOX[XUL ILQDQFLDUH GDWRUDWH SHULRDGHORU VXFFHVLYH 5HJXOLOH DFWXDOL] ULL

h (h<1) avem:

- 1 leu plasat n t=0 este echivalent cu (1+rh) lungime h; - 1 leu plasat n t=0 este echivalent cu (1+rh) -k
vQ FRQVHFLQ  YDORDUHD DFWXDOL]DW

OHL OD VIkULWXO XQHL SHULRDGH GH

OHL OD VIkULWXO D

k perioade;
SHULRDG HVWH

D  OHX GLVSRQLELO vQ D

k-D

(1+rh) lei. S-D FRQVLGHUDW F

UDWD GH DFWXDOL]DUH SH XQLWDWHD GH WLPS HVWH

r.

Criteriul J devine: J=

bk(Xk-1,uk)(1+rh) h + m(XT/h)(1+rh)
-1/h -t/h

-k

-T/h

sau
-T/h

J=[ b(X(0),u(1),1)(1+rh)
UHVSHFWLY P

+...+b(X(t-h),u(t),t)(1+rh) ]h+...+m(X(T)) (1+rh)

YDORDUHD FkWLJXOXL UHDOL]DW vQ

unde bk

GHVHPQHD]

SHULRDGD

k,

(respectiv T/h HYDOXDW

OD VIkULWXO DFHVWHL SHULRDGH 3ULQ PXOWLSOLFDUHD FX IDFWRUXO GH YDORDUHD DFWXDO D FkWLJXULORU

actualizare, criteriul J H[SULP

RE LQHP PRGHOXO

Particulariznd valorile OXL K VH RE LQ GLIHULWH FODVH GH PRGHOH GLQDPLFH $VWIHO pentru h=1
GLQDPLF GLVFUHW SHQWUX

h<1 modelul dinamic este

discretizat, iar pentru h0, modelul dinamic devine continuu.

5.2. Modelul dinamic discret (MDd)

Este definit pentru perioade elementare de lungime h=1. El se scrie astfel: Max J =

bt(Xt-1, ut)(1+r) + m(XT)(1+r) [3] [4]

-t

-T

Xt-Xt-1 = at(Xt-1,ut) , t=1,...,T ut U , X0=

0'G SRVHG R VWUXFWXU SDUWLFXODU 

este un model cu 2T variabile (u1,u2,...,uT L X1,X2,...,XT);

HFXD LD GH WUDQ]L LH >@ SXQH vQ UHOD LH vQ ILHFDUH SHULRDG

HOHPHQWDU  QXPDL 

variabile: Xt-1 L Xt;

UHOD LD >@ VH UHIHU  vQ ILHFDUH SHULRDG  GRDU OD YDULDELOD GH FRQWURO FRUHVSXQ] WRDUH SHULRDGH VROX LD RSWLPDO D

i respective;
HVWH FRQVWLWXLW GLQ WUDLHFWRULD RSWLPDO

0'G

X*=(X*1,X*2,...,X*T) L GLQ SROLWLFD RSWLPDO u*=(u*1,u*2,...,u*T);

vQ PRGHOHOH GLQDPLFH GLVFUHWH FRQFUHWH PDL SRW LQWHUYHQL XUP WRDUHOH FRPSOLFD LL

- VH LPSXQ UHVWULF LL DVXSUD VW

ULL ILQDOH D IHQRPHQXOXL PRGHODW

fT(XT) 0;

- se LPSXQ UHVWULF LL DVXSUD VW ULL IHQRPHQXOXL n toate perioadele: ft(Xt) 0, t=1,...,T. - HFXD LD GH WUDQ]L LH HVWH R HFXD LH UHWDUGDW
FDUH H[SULP 

Xt-Xt-1=at(Xt-1,Xt-2,...,Xt-s, ut),
IHQRPHQH GH PHPRULH LQGXFkQG R GHSHQGHQ PDL SXWHUQLF ID GH

trecut.
([LVW QXPHURDVH IHQRPHQH L SURFHVH HFRQRPLFH FDUH VH PRGHOHD] FX DMXWRUXO

(MDd). Cele mai frecvente sunt procesele de aprovizionare-stocare, procese de

RUGRQDQ DUH LQYHVWL LL HWF

5.3. Modelul dinamic continuu (MDc)

6H RE LQH GLQ PRGHOXO GLQDPLF JHQHUDO SULQ WUHFHUH OD OLPLW  I FkQG UHOD LLORU VFULVH VXE IRUP IXQF LRQDO 

h0, a

lim (1+rh)

-t/h

=e

-rt

lim [X(t) - X(t-h)]/h = dX(t)/dt.

'HULYDWD YDULDELOHL GH VWDUH VH QRWHD]

FX

X( t ) . Vectorul X ( t)

HVWH GH DFHHDL

dimensiune cu vectorul X(t) 2E LQHP DVWIHO PRGHOXO 0'F

Max J =

b(X(t),u(t),t)e dt + m(X(T))e
X( t )=a(X(t),u(t),t)

-rt

-rT

u(t) U, X(0)=
6ROX LD PRGHOXOXL VH FDXW SULQWUH IXQF LLOH

X : [0,T]R
(VWH XQ PRGHO SDUWLFXODU GH

X  >7@

R .
m
PRGHOHOH GLQDPLFH

GLPHQVLXQH

ILQLW 

8QHRUL

FRQWLQXL SUH]LQW

RUL]RQW LQILQLW &kQG QX H[LVW

ULVFXO DPELJXLW

LL SXWHP UHQXQ D OD

PHQ LRQDUHD YDULDELOHL W 0RGHOXO GHYLQH vQ DFHVW FD]

Max J =

-rt -rT b(X,u,t)e dt + m(X(T))e

X( t ) = a(X,u,t)

uU , X(0) = .

(MDc) este utilizat pentru modelarea proceselor controlabile n orice moment,

DGLF vQ WLPS UHDO 'DWRULW FRPRGLW LL IRORVLULL FDOFXOXOXL GLIHUHQ LDO L LQWHJUDO PRGHOXO FRQWLQXX VH SUHWHD] WLPSXOXL HVWH DUELWUDU  ' P vQ FRQWLQXDUH XQ H[HPSOX OD DQDOL]D IHQRPHQHORU HFRQRPLFH SHQWUX FDUH VFDUD

de utilizare a modelelor dinamice.

5.4. Gestiunea unei resurse naturale

&RQVLGHU P R vQWUHSULQGHUH FDUH H[SORDWHD] PLQ R UHVXUV QDWXUDO GH H[HPSOX R

GH PHWDOH SUH LRDVH VDX R H[SORDWDUH IRUHVWLHU  )LH RUL]RQWXO > H[SORDWDW HVWH LPHGLDW YkQGXW SH R SLD

0,T]. Cantitatea

GH UHVXUV

XQGH vQWUHSULQGHUHD GH LQH R

SR]L LH

GH

PRQRSRO )DFHP

DEVWUDF LH SHQWUX

PRPHQW

GH

FRVWXULOH GH H[SORDWDUH

QWUHSULQGHUHD GRUHWH V vQFkW V

SURJUDPH]H

H[SORDWDUHD UHVXUVHL SH RUL]RQWXO GDW DVWIHO FX R UDW

PD[LPL]H]H VXPD EHQHILFLLORU DFWXDOL]DW

r.
HVWH UHSUH]HQWDW SULQ

a) MDd. Activitatea ntreprinderii pe perioada t=(t-1,t] variabilele: nivelul resursei la nceputul perioadei t, Xt-1;
FDQWLWDWHD H[WUDV vQ SHULRDGD

t, ut (ut0);
H[SORDWDW 

SUH XO SLH LL vQ SHULRDGD

t, care deSLQGH GH FDQWLWDWHD GH UHVXUV

p(ut).

Q WHRULD PLFURHFRQRPLF  IXQF LD FDUH G

SUH XO XQXL EXQ vQ IXQF LH GH FDQWLWDWHD GLQ DFHO EXQ VH HVWH YRUED GHVSUH XQ EXQ QRUPDO DGLF XQ EXQ OD

QXPHWH IXQF LD FHUHULL LQYHUVH 3UHVXSXQHP F FDUH SUH XO VFDGH SH P VXU

FH FDQWLWDWHD GH EXQ FUHWH $FHDVWD UHYLQH OD FRQGL LD FD GHULYDWD GH ILH QHJDWLY 

RUGLQXO , D IXQF LHL FHUHULL LQYHUVH V

p<0);

QLYHOXO LQL LDO DO UHVXUVHL HVWH FXQRVFXW 

X0=

,PSXQHP GHDVHPHQHD R FRQGL LH

L DVXSUD VW ULL ILQDOH FDQWLWDWHD GH UHVXUV WUHEXLH V GHS HDVF XQ QLYHO PLQLP GDW 

OD VIkULWXO RUL]RQWXOXL GH WLPS 

XT)

s), evitnd epuizarea resursei.

-t

0RGHOXO GH JHVWLXQH RSWLPDO

D UHVXUVHL QDWXUDOH HVWH

Max J =

p(ut)ut(1+r)

Xt-Xt-1 = ut ut 0 control (m=1

ILQDO 

, XTs , X0 = .
GH VWDUH  FX

$FHVWD HVWH XQ 0'G FDUDFWHUL]DW GH R YDULDELO R PXO LPH GH GHFL]LL DGPLVLELOH

p=1

R YDULDELO

GH

HJDO

R+

L R UHVWULF LH GH VWDUH

XTs.

b) MDc.

Fie u(t)
FDQWLWDWHD H[WUDV RULFH PRPHQW V

IXQF LD

FDUH

P VRDU

YLWH]D

GH

H[SORDWDUH

UHVXUVHL

UHVSHFWLY

vQ XQLWDWHD GH WLPS OD GDWD

t 3UHVXSXQHP F

vQWUHSULQGHUHD SRDWH vQ HVWH

DMXVWH]H YLWH]D GH H[SORDWDUH 0RGHOXO GH JHVWLXQH RSWLPDO

Max J =

p(u)ue dt
X = -u

-rt

u0, X(T)s, X(0)= .

5H]ROYDUHD PRGHOXOXL GHWHUPLQ SUHVXSXQH F ULWPXO vQ FDUH YD IL H[SORDWDW UHVXUVD 6H

vQWUHSULQGHUHD FXQRDWH UDWD GH DFWXDOL]DUH

r

QLYHOXO LQL LDO [

, minimul

ILQDO V L IXQF LD FHUHULL LQYHUVH * VLUHD HIHFWLY

p(u).

D VROX LHL VH IDFH FX DMXWRUXO SULQFLSLXOXL PD[LPXOXL DO OXL

3RQWUHDJKLQ FH YD IL SUH]HQWDW vQ SDUDJUDIXO XUP WRU

5.5. Principiul maximului

3ULQFLSLXO OXL 3RQWUHDJKLQ IXUQL]HD] FRQGL LLOH QHFHVDUH GH RSWLP vQ P

odelele

GLQDPLFH 9RP VWDELOL DFHVWH FRQGL LL SHQWUX 0'G LDU DSRL SULQ WUHFHUH OD OLPLW 

pentru (MDc).
3HQWUX VLPSOLILFDUHD VFULHULL YRP DQDOL]D FD]XO FkQG H[LVW R VLQJXU PXO LPHD YDULDELO

de stare (p=1 admisibile U

VLQJXU

YDULDELO

GH

FRQWURO

m=

LDU

deciziilor

VH H[WLQGH OD WRW VSD LXO

DFHDVWD vQVHPQkQG GH IDSW F

QX LPSXQHP

UHVWULF LL DVXSUD YDULDELOHL GH GHFL]LH 1H UHIHULP GHFL OD PRGHOXO 0'G (FXD LLORU GH WUDQ]L LH >@ vQ QXP U GH

T

OL VH DVRFLD]

PXOWLSOLFDWRULL /DJUDQJH

. Fie t

multiplicatorul asociat celei de a t-D

UHSUH]LQW YDORDUHD DFWXDOL]DW

UHVWULF LH 1RW P

t = (1+r) t. Variabila t
t
3HQWUX FRHUHQ  YRP

D PXOWLSOLFDWRUXOXL

t la data t

nota cu 0

PXOWLSOLFDWRUXO DVRFLDW UHVWULF LHL

X0= LUXO GH YDULDELOH 0,1,2

,,T se numesc variabile adjuncte ale modelului dinamic. Scriem acum lagrangeanul problemei (MDd): L ()= [bt(Xt-1,ut)-t(Xt-Xt-1-at(Xt-1,ut))](1+r) +m(XT)(1+r) - 0(X0- ).
-t -T
t

&RQGL LLOH

GH

RSWLPDOLWDWH

VH

GHGXF

GLQ

DQXODUHD

GHULYDWHORU

SDU LDOH

DOH

lagrangeanului:

derivatele n raport cu ut, t=1,...,T [bt/ut+ tat/ut ](1+r) = 0

-t

derivatele n raport cu Xt-1, pentru t=2,...,T

bt/Xt-1+t(1+at/Xt-1)](1+r) -t-1(1+r)
-t

-(t-1)

=0

derivata n raport cu X0: [bt/X0 + 1(1+a1/X0)](1+r) - 0= 0

-1

derivata n raport cu XT: [- T+m(XT)](1+r) = 0

-T
'XS VLPSOLILF ULOH SRVLELOH VH RE LQ UHOD LLOH

bt/ut +tat/ut = 0 t - (1+r) t-1 = - [bt/Xt-1+tat/Xt-1], t=1,2,...,T t = m(XT)

2ULFH VROX LH RSWLPDO 

X*,u*

D PRGHOXOXL 0'G HVWH DVRFLDW

XQXL LU GH

* * * variabile adjuncte * = (* 0 , 1 ,..., t ,..., T ) DVWIHO vQFkW V

ILH UHDOL]DWH FRQGL LLOH

X t* X t* 1 = at ( X t* 1 , ut* ), t = 1,2,..., T [I] * X0 =

* * * bt ( X t* 1 , ut* ) * * at ( X t 1 , ut ) + r = + ( 1 ) [ ] t t 1 t X t 1 X t 1 [II] * = m , X * T T

( )

[III]

bt ( X t* 1 , ut* ) a ( X * , u * ) + t* t t 1 t = 0 ut ut

3HQWUX D LQWHUSUHWD PDL XRU DFHVWH UHOD LL VH GHILQHWH R QRX

IXQF LH QXPLW

hamiltonian.
6H QXPHWH KDPLOWRQLDQXO PRGHOXOXL GLQDPLF 0'G IXQF LD

Ht GHILQLW

SULQ

Ht = bt(Xt-1,ut)+tat(Xt-1,ut).
5HOD LLOH GH PDL VXV GHYLQ

* H t* * = , t = 1,2,..., T X X t t 1 t [I] X* = 0 * H t* * , t = 1,2,..., T t (1 + r ) t 1 = X t 1 [II] * = m X * T T

( )

[III]

H t* = 0, ut

t = 1,2,..., T

6H REVHUY

vQ XOWLPHOH UHOD LL F

YDULDELOHOH

Xt L t MRDF

XQ URO VLPHWULF 

variabila de stare Xt HVWH VROX LD XQHL HFXD LL GH WUDQ]L LH FX FRQGL LH LQL LDO
YDULDELOD DGMXQFW FRQGL LH ILQDO  DO GRLOHD WHUPHQ DO XQHL HFXD LL GH WUDQ]L LH DO XQHLD GLQ FHOH GRX VHPQ VFKLPEDW GHULYDWD SDU LDO

HVWH GHDVHPHQHD VROX LD XQHL HFXD LL GH WUDQ]L LH GDU FX

YDULDELOH HVWH FX

D KDPLOWRQLDQXOXL vQ UDSRUW FX FHDODOW  vQ IXQF LH GH

FRQIRUP FX >,,,@ YDULDELOD GH FRQWURO VH H[SULP

Xt-1 L t;

vQ DFHVWH FRQGL LL VLVWHPXO FRQVWLWXLW GLQ FHOH GRX VLVWHP FXSODW DGLF

HFXD LL GH WUDQ]L LH HVWH XQ

PHPEUXO VWkQJ DO ILHF UHLD GHSLQGH GH

Xt-1, respectiv t.
DVRFLD L SXQFW HFXD LHL GH GH

Variabilele adjuncte sunt multiplicatorii Kuhn-7XFNHU

WUDQ]L LH (OH DX R LQWHUSUHWDUH HFRQRPLF DVHP Q WRDUH

GLQ

YHGHUH

PDUJLQDOLVW DUDW

YDULD LD VXIHULW

GH FULWHULXO GH RSWLP GDWRUDW

YDULD LHL YDULDELOHL GH OD DFHD VXIHU R D

stare. Variabila s*
GDW SHUWXUED LH GH R

HVWH YDORDUHD OD GDWD V DWULEXLW

VW ULL VLVWHPXOXL DWLQV YDULDELOD GH VWDUH YDULD LH D VXPHL

SULQ WUDLHFWRULD RSWLP  &X DOWH FXYLQWH GDF XQLWDWH FX OD GDWD

X*s

s

YD

UH]XOWD

DFWXDOL]DWH

FkWLJXULORU HJDO

s*.

2EVHUYD LH

&RQVLGHU P PRGHOXO 0'G V RE LQXW SULQ WUXQFKLHUHD PRGHOXOXL 0'G SHQWUX

a restrnge orizontul de timp la intervalul [s,T], pornind la data s GLQ SR]L LD LQL LDO . Modelul devine:
t = s +1

[ bt(Xt-1,ut)(1+r)

-(t+s)

+ m(XT)(1+r)

-(T-s)

Xt-Xt-1 = at(Xt-1,ut) , t=s,s+1,...,T ut U , Xs =

1RW P FX >

X*t([,s),t=s,...,T] traiectoria sistemului,

UHJ VLP PRGHOXO 0'G DGLF  SULQFLSLXO PD[LPXOXL

[*t([,s), t=s,...,T]
L

YDULDELOHOH DGMXQFWH L REVHUY F SHQWUX

J*([,s) valoarea criteriului de optim a modelului trunchiat. Se X*t=X*t([,0), *t=*t([,0)

WUXQFKLDW VH RE LQ F WUDLHFWRULLOH RSWLPDOH

s=0

J*=J*([,0)
FRQGL LLOH GH

$SOLFkQG

PRGHOXOXL DUDW

H[DFW DOH

RSWLPDOLWDWH

DQWHULRDUH

$FHDVWD

PRGHOHORU 0'G L 0'G[ SR]L LD LQL LDO

,s) sunt acelHDL SH LQWHUYDOXO >s,T@ ELQHvQ HOHV QXPDL GDF

OD GDWD

D PRGHOXOXL WUXQFKLDW OD GDWD V FRLQFLGH FX SR]L LD DWLQV D PRGHOXOXL 0'G DGLF 

s de

WUDLHFWRULD RSWLPDO

X*t(X*s,s) = X*t, *t (X*s,s) = *t pentru

PRGHOH HVWH

RULFH V L W DVWIHO vQFkW

s<t 5HOD LD GLQWUH YDORULOH FULWHULLORU FHORU GRX

s t =1

-t -s J* = bt(X*t-1,u*t)(1+r) + J*(X*s,s)(1+r) .

$FHDVW

UHOD LH H[SULP

principiul de optimalitate al lui Bellman,

PD[LPXOXL FRQGXFH OD

FDUH VW

OD

baza unui demers alternativ al modelelor dinamice numit SURJUDPDUH GLQDPLF .

Q FD]XO FRQWLQXX SULQFLSLXO XUP WRDUHOH FRQGL LL

necesare de optimalitate:
* H * (t ) , t [0, T ] X (t ) = [I] X * ( 0) = * H * (t ) * t r t = ( ) ( ) , t = 1,2,..., T [II] X * (T ) = m , X * (T )

H * (t ) = 0, [III] u

t = 1,2,..., T

$FHVWH FRQGL LL VXQW L VXILFLHQWH GDF

KDPLOWRQLDQXO

HVWH R IXQF LH FRQFDY GH

n X L u

Q VFULHUHD FRQGL LLORU V D SUHVXSXV F

0'F HVWH GHILQLW FD R SUREOHP

GLPHQVLXQH ILQLW GLVFUHWL]DW F WUH

L F FHOH

DUH ORF FRQYHUJHQ D FRQGL LLORU GH RSWLPDOLWDWH DOH PRGHOXOXL DOH PRGHOXOXL FRQWLQXX $FHVWH LSRWH]H VXQW VDWLVI FXWH vQ

majoritatea modelelor economice. Adesea, modelele economice sunt caracterizate de mai multe variabile de stare
L GH FRQWURO QX QXPDL GH FkWH XQD DD FXP DP SUHVXSXV SkQ UHVWULF LL GH OHJ WXU vQWUH DFHVWHD 3ULQFLSLLOH HQXQ DWH PDL VXV DFXP FD L GH U PkQ YDODELOH

vQORFXLQG SURGXVXO GH QXPHUH SULQ SURGXVXO VFDODU L GHULYDWHOH IXQF LLORU SULQ YHFWRUL GH JUDGLHQW )RUPXO P GHFL SULQFLSLXO PD[LPXOXL vQWU L

-un cadru mai larg, n care X(t) C(X(T))0, unde C

u(t) sunt vectori n R

 LDU UHVWULF LD GH VWDUH ILQDO L FRQYH[ GH OD

HVWH GH WLSXO

HVWH R IXQF LH GLIHUHQ LDELO

R la R . Hamiltonianul, n acest caz este:

i i

H = b(X(t),u(t),t) + (t)a (X(t),u(t),t),

i

unde ai () este a i-D

0'F VH HQXQ &RQGL LD

FRPSRQHQW

D DSOLFD LHL

a. Principiul maximului pentru modelul

DVWIHO QHFHVDU SHQWUX FD

[X*(t), u*(t)]

V )

ILH

VROX LH  GHILQLW

RSWLPDO SH >

PRGHOXOXL 0'F HVWH V

H[LVWH IXQF LD YHFWRULDO

*(t)R

0,T@ i

un vector J*R , astfel nct pentru orice t[0,T] V

DYHP

[1] H(X*(t),u*(t),)*(t),t)t H(X*(t),u(t),)*(t),t), uU [2] [3] [4]

) )

(t) = r)*(t) -x H*
*

(T) = m(X*(T)) + J*xC(X*(T)) 0 , J*C(X*(T)) = 0

J t

&RQGL LLOH >@ L >@ VH QXPHVF

condL LL

GH WUDQVYHUVDOLWDWH

ale problemei de

control optimal.

5.6. Model dinamic de dezvoltare a ntreprinderii

6FRSXO DFHVWXL SDUDJUDI HVWH V LOXVWUH]H WHKQLFLOH XWLOL]DWH vQ PRGHODUHD

IHQRPHQHORU L SURFHVHORU HFRQRPLFH DWXQFL FkQG VH DO DFHVWRUD DGLF

LQH VHDPD GH DVSHFWXO G

inamic

DWXQFL FkQG YDULDELOHOH LVDX SDUDPHWULL PRGHOXOXL VXQW IXQF LL GH HYROX LD IHQRPHQXOXL VDX SUHFHVXOXL HFRQRPLF SH XQ

WLPS LDU PRGHOHOH LOXVWUHD]

orizont de timp dat.

Vom construi un model dinamic al ntreprinderii, particularizat la deciziile de

GH]YROWDUH D DFWLYLW LSRWH]H GH ED]  LL VDOH QFHSHP VSHFLILFDUHD PRGHOXOXL SULQ SUH]HQWDUHD XQRU

Ipoteza 1 (I1

9LD D vQWUHSULQGHULL VH GHUXOHD]

SH R VXFFHVLXQH GH

T perioade

DQL WULPHVWUH HWF (VWH LSRWH]D GH ED] SURFHVHOH HFRQRPLFH FDUH VH GHVI RDU

vQ PRGHOHOH GLQDPLFH )HQRPHQHOH L

vQ vQWUHSULQGHUH QX VXQW DERUGDWH VWDWLF OD

XQ DQXPLW PRPHQW FL vQ HYROX LD ORU SH XQ RUL]RQW GH WLPS IL[DW VHFYHQ LRQDOL]DW

n T perioade.

Ipoteza 2 (I2): ntreprinderea dispune de un centru de decizie unic (de exemplu,

$GXQDUHD *HQHUDO D $F LRQDULORU $FHDVW LSRWH] IL[HD] PRGXO GH DERUGDUH D SUREOHPHORU GHFL]LRQDOH OHJDWH GH GH]YROWDUHD VD Q XUPD GLVFXW ULL L DSURE ULL ELODQ XOXL FRQWDELO vQ $*$ VH GHFLGH DVXSUD GH]YROW ULL

ntreprinderii pe

SHULRDGD XUP WRDUH 'DF

V DU UHQXQ D OD R DVWIHO GH LSRWH]  DU LQWHUYHQL SUREOHPH

de decizie multinivel care presupun o coordonare a decizziilor pe nivele

GHFL]LRQDOH ,SRWH]D QX HVWH UHVWULFWLY  H[LVWkQG PHWRGH L WHKQLFL GH RSWLPL]DU

ea
DO

deciziilor multinivel.

Ipoteza 3 (I3
vQWUHSULQGHULL YLLWRUXO HVWH L

6H

UHIHU

OD

FRQGL LLOH F

DVXSUD VXQWHP FX

HYROX LHL vQ

PHGLXOXL GH

HFRQRPLF

DQXPH

SUHVXSXQH

FRQGL LL

FHUWLWXGLQH F vQWUHSULQ

GHFL

SHUIHFW

FXQRVFXW

PSUHXQ

I1

DFHDVWD

DUDW

derea

DF LRQHD]

vQ FRQGLWLL GH

SUHYL]LXQH SHUIHFW

 DGLF  GH H[HPSOX HD FXQRDWH

H[DFW FDUH YD IL HYROX LD FRVWXULORU GH SURGXF LH D FHUHULL GH SURGXVH HWF SH

ntregul orizont de timp considerat [0,T@ oarecaUH

IRDUWH PXOW PRGHOXO ,QFHUWLWXGLQHD L

$FHDVW

LSRWH]

HVWH UHVWULFWLY

vQWU

-o

P VXU  GDU OXDUHD vQ FRQVLGHUDUH H[SOLFLW

D LQFHUWLWXGLQLL DU FRPSOLFD

ULVFXO OHJDW GH DFHDVWD DX IRVW DERUGDWH

separat n capitolul III consacrat acestor tipuri de modele.

Ipoteza 4 (I4 6H UHIHU

F IXQF LH GH XWLOLWDWH

OD FRPSRUWDPHQWXO DF LRQDULORU vQWUHSULQGHULL L SUHVXSXQH UD LRQDO  XUP ULQG V PD[LPL]H]H R

DFHWLD DX XQ FRPSRUWDPHQW HFRQRPLF

U=U(v1,...,vT), unde vt

HVWH YHQLWXO QHW UHDOL]DW GH DF LRQDUL vQ

perioada t, t=1,2,...,T ,SRWH]D IL[HD] {Max U}

vQVHDPQ DF LRQDUL

 GHFL IXQF LD RELHFWLY D PRGHOXOXL FDUH YD IL DSDU  $YkQG vQ YHGHUH L DEVROXW

HYLGHQW VXE GLYHUVH UHVWULF LL FDUH SRW V F SUHVXSXQHP R FRRUGRQDUH SHUIHFW JHVWLRQDUL PDQDJHUL ERJ 0DQDJHULL

I2,
GH

L R WUDQVSDUHQ VH DQJDMHD]  SULQ

vQWUH

FRQWUDFWXO

PDQDJHPHQW V GH GH]YROWDUH

VSRUHDVF

LD DF LRQDULORU (L YRU SUH]HQWD DFHVWRUD VWUDWHJLD FH GHFXUJ GLQ HD GDU GHFL]LD ILQDO

D vQWUHSULQGHULL L GHFL]LLOH

DSDU LQH DF LRQDULORU FDUH XUP UHVF QX R PD[LPL]DUH D YHQLWXOXL QHW W PD[LPL]DUH D XWLOLW LL SH FDUH R RE

otal, ci o

in cu aceste venituri.
OD IDSWXO F PRG GH LPSR]LWHOH SO WLWH GH FDOFXO QX VFKLPE

Ipoteza 5 (I5

1X H[LVW

QLFL R GLVWRUVLXQH vQWUH ILVFDOLWDWHD vQWUHSULQGHULL L D

DF LRQDULORU FD SHUVRDQH IL]LFH ,SRWH]D VH UHIHU

ntreprindere sXQW

ILH

LPSR]LWH

GLUHFWH

DO

F URU

FRPSRUWDPHQWXO DF LRQDULORU ILH LPSR]LWH LQGLUHFWH SH YHQLWXULOH GLQ GLYLGHQGH DOH DF LRQDULORU FDUH LQIOXHQ HD] DFHVW FRPSRUWDPHQW SLD GH FDSLWDO V SHUIHFW  $FHDVWD SUHVXSXQH F

Ipoteza 6 (I6
DJHQ L

([LVW

vQWUHSULQGHUHD SRDWH OD RULFH GDW HFRQRPLFL VXE IRUPD

t

LD FX vPSUXPXW VDX V RULFH VXPH GH

vPSUXPXWH DO L OD R UDW D

SODVDPHQWHORU

EDQL

dobnzii r FXQRVFXW

GLQDLQWH

Acestea au fost doar cteva probleme legate de specificarea modelului dinamic

DO vQWUHSULQGHULL 3HQWUX R VSHFLILFDUH FRPSOHW  WUHEXLH LQWURGXVH L DOWH QR LXQL HFRQRPLFH L FRQWDELOH $FHVW IDSW YD GHWHUPLQD FD SH P VXUD FRQVWUXLULL PRGHOXOXL V LQWURGXFHP L DOWH LSRWH]H

Conform cu considera iile f cute la modelele dinamice, trebuie definite

RUL]RQWXO GH WLPS YDULDELOHOH GH VWDUH L GH FRPDQG L HFXD LD GH WUDQ]L LH D VW ULORU

Definirea elementelor modelului dinamic

Orizontul de timp al modelului poate fi reprezentat grafic astfel: |---|---|-----------------|---|-------------------------|--> 0 1 2 perioada t HVWH GHWHUPLQDW ...
GH FDUH P VRDU VXPD RSHUD LLORU UHDOL]DWH vQ

t-1 t

...

T
(YROX LD vQWUHSULQGHULL SH

Perioada t a orizontului va fi intervalul (t-1,t@ variabile "flux", notate prin vectorul ut perioada t; variabile "stoc", reprezentate prin vectorul Et

FDUH FDUDFWHUL]HD]

HIHFWXO FXPXODW

OD XQ PRPHQW GDW DO DF LXQLORU vQWUHSULQGHULL vQ WRDWH SHULRDGHOH WUHFXWH

VariaELOHOH
YDULDELOHOH VWRF

IOX[ FRUHVSXQG GHFL]LLORU FH WUHEXLH OXDWH vQ RULFH SHULRDG VWDUHD vQWUHSULQGHULL OD RULFH GDW PRPHQW W

t, iar
6WDUHD

UHSUH]LQW

ntreprinderii la data t, Et

HVWH GHWHUPLQDW

GH VWDUHD OD GDWD

t-1, Et-1

L GH GHFL]LLOH

ut

luate n perioada t=(t-1,t] 3XWHP VFULH F

1,ut)

HFXD LD GH WUDQ]L LH D VW ULORU HVWH GDF

Et=Ft(EtWUDQ]L LD

>@ $VWIHO HYROX LD vQWUHSULQGHULL YD IL SHUIHFW GHWHUPLQDW

VH FXQRVF

VWDUHD LQL LDO 

E0

LUXO GH GHFL]LL >

u1,u2,...,uT@

L IXQF LLOH

Ft

FDUH GHWHUPLQ

de la o stare la alta.
'HILQLP DFXP SULQFLSDOHOH YDULDELOH IOX[ DVRFLDWH DFWLYLW LL vQWUHSULQGHULL SH

perioada t:

1. Cifra de afaceri: CAt 2. Cheltuielile de exploatare: CHt

 &KHOWXLHOLOH GH LQYHVWL LL

It

4. Cheltuielile financiare: CFt 5. Impozitul pe profit: IMt 6. Amortizarea: At

 9DULD LD GDWRULLORU

xt kt

 &UHWHUHD FDSLWDOXOXL

'HVLJXU F VXILFLHQWH 3H

VH SRW GHILQL L DOWH YDULDELOH 3HQWUX QHYRLOH QRDVWUH DFHVWHD VXQW ORU FDOFXO P XQ QXP U GH LQGLFDWRUL GH UH]XOWDWH GHF

ED]D

i alte

variabile:
 0DUMD EUXW GH H[SORDWDUH

MBt = CAt - CHt

10. Profitul brut: PBt = MBt - CFt 11. Profitul "net" : PNt = PBt - At
 3URILWXO GXS LPSR]LWDUH

Pt = PNt - WtPNt = (1- Wt) PNt

unde Wt este rata de impozit. Putem considera, simplifiFkQG F

SURILW QX VH PRGLILF 3URILWXO GXS W

t=W

 RULFDUH DU IL W DGLF

UDWD GH LPSR]LW SH

SH RUL]RQWXO GH PRGHODUH L

Pt = (1-W) PNt .
GHVWLQD LL R SDUWH 3'

LPSR]LWDUH HVWH UHSDUWL]DW SH GRX DF LRQDULORU LDU UHVWXO

t) se

GLVWULEXLH FD GLYLGHQGH

HVWH UH LQXW GH vQWUHSULQGHUH SHQWUX

dezvoltare (PRt). Avem egalitatea: Pt = PDt + PRt [2].

 (FRQRPLVLUHD EUXW 

et = PRt + At

14. Fluxul de numerar (cash-flow): FNt = Pt + At

Q SUH]HQWDUHD I FXW DX IRVW QHJOLMDWH VWRFXULOH GH PDWHULL SULPH L SURG

use

ILQLWH WUH]RUHULD SDUWLFLSDUHD OD DOWH VRFLHW 'HILQLP DFXP YDULDELOHOH VWRF FH

L SOXV YDORULOH L YkQ] ULOH GH LPRELOL] UL

FDUDFWHUL]HD]

DFWLYLWDWHD

vQWUHSULQGHULL

3HQWUX DFHDVWD SRUQLP GH OD ELODQ XO FRQWDELO DO vQWUHSULQGHULL FDUH vQ PRG VF

hematic,

la data t, cuprinde n partea de Activ, activele imobilizate, iar n cea de Pasiv,

IRQGXULOH SURSULL L GDWRULLOH &RQVLGHUkQG SH GH R SDUWH DFWLYXO QHW FRQWDELO $1 LDU SH GH DOW IXQGDPHQWDO  SDUWH IRQGXULOH SURSULL )3  L GDWRULLOH ;

t),

t),

scriem rela LD

FRQWDELO

ANt = FPt + Xt [3]. Variabilele stoc ANt, FPt, Xt sunt variabilele de vQWUHSULQGHUHD (YROX LD ORU

VWDUH HVHQ LDOH DOH VLVWHPXOXL GLQDPLF SH FDUH O UHSUH]LQW HVWH GHVFULV GH XQHOH UHOD LL FRQWDELOH FODVLFH

YDULD LD DFWLYXOXL QHW FRQWDELO vQWUH GRX

SHULRDGH HVWH HJDO

FX GLIHUHQ D GLQWUH

LQYHVWL LL L DPRUWL]DUH

ANt - ANt-1 = It - At [4].

SHULRDGH HVWH HJDO FX FUHWHUHD FDSLWDOXOXL

YDULD LD IRQGXULORU SURSULL vQWUH GRX

SOXV FUHWHUHD SURILWXULORU UH LQXWH SHQWUX GH]YRO

tare:
HJDO FX VXPD vPSUXPXWDW

FPt - FPt-1 = kt + PRt [5].

-YDULD LD

GDWRULLORU

vQWUH

GRX [

SHULRDGH

HVWH

UDPEXUVDW

VDX SODVDW  GDF

t<0):

Xt - Xt-1 = xt

[6]. t-1 L t, devine:

5HOD LD >@ VFULV

SHQWUX GRX

PRPHQWH VXFFHVLYH

ANt-1 = FPt-1 + Xt-1 ANt = FPt + Xt

sau: (ANt - ANt-1) = (FPt - FPt-1) + (Xt - Xt-1)

LQkQG FRQW GH UHOD LLOH >@ >@ L >@ RE LQHP

It - At = kt + PRt +xt

[7]


sau It = kt + At + PRt + xt L FXP PRt + At = et  UH]XOW It = kt +xt + et [8]

5HOD LLOH >@ L >@ VXQW UHOD LL XWLOL]DUH UHVXUVH

L H[SULP

IDSWXO F

LQYHVWL LLOH 

realizate n perioada t (It mprumut (xt veniWXO

VXQW

ILQDQ DWH $FHDVW

SULQ

FUHWHUHD

FDSLWDOXOXL

SURSULX

kt),

L HFRQRPLVLUH 

et

UHOD LH VH YD VFULH GLIHULW GDF

GHILQLP

QHW DO DF LRQDULORU FD GLIHUHQ

vQWUH SURILWXO GLVWULEXLW L FUHWHUHD FDSLWDOXOXL

vt = PDt - kt [9].
'LQ UHOD LLOH DQWHULRDUH GHGXFHP F 

Pt = (1-W)PNt = (1-W)(PBt - At) = (1-W)(MBt - CFt - At).

Q DFHODL WLPS LQkQG FRQW GH >@ GHGXFHP HJDO

itatea:

PDt +PRt = (1-W)(MBt - CFt - At) [10]. Dar, PRt + At = et DGLF PRt = et - At, L vQORFXLQG vQ >@ UH]XOW (1-W)(MBt - CFt - At) = PDt + et - At , sau: et = (1- W)(MBt - CFt - At) - PDt + At [11].
5HOD LD >@ GHYLQH 

It=kt+xt+et=kt+xt+[(1-W)(MBt-CFt-At)-PDt+At],
DGLF 

It+(PDt-kt) = xt + (1-W)(MBt-CFt)+ WAt

L LQkQG FRQW GH >@ UH]XOW 

It + vt = xt + (1-W)(MBt - CFt) + WAt [8].

6XE DFHDVW VXPD LQYHVWLW SURILWXO GXS IRUP  UHOD LD XWLOL]DUH UHVXUVH

DUDW

F  vQ ILHFDUH SHULRDG

t,
GH

SOXV YHQLWXO DF LRQDULORU HVWH HJDO LPSR]LWDUH L WHUPHQXO

FX VXPD GLQWUH YDULD LD GDWRULLORU GLQWUH DPRUWL]DUH L UDWD

At

SURGXVXO

LPSR]LWDUH FDUH VH LQWHUSUHWHD]

FD XQ FUHGLW GH LPSR]LW OHJDW GH OXDUHD vQ FDOFXO

fLVFDO D DPRUWL]

ULL GHRVHELW  LQYHVWL LLOH SRVLELOH D IL UHDOL]DWH GH vQWUHSULQGHUH L UHQWDELOLWDWH GDU ILQDQ DUHD ORU HVWH GHFLV V YHGHP GDF  vQ DQXPLWH FRQGL LL

$SDUH R SUREOHP WUHEXLH MXGHFDWH GXS GXS FULWHULL SURSULL

FULWHULL GH HILFLHQ DF LRQDULORU

7UHEXLH

ntreprinderea poate lua decizii privind dezvoltarea sa viitoare pe criterii pur

HFRQRPLFH L WHKQLFH LQGHSHQGHQW GH ILQDQ DUHD L GLVWULEXLUHD GLYLGHQGHORU DGLF LQGHSHQGHQW FRUHFW GH FULWHULLOH GH OD SURSULL LSRWH]D DF LRQDULORU , GLVFXWDW 3UREOHPD DQWHULRU ERJ SRDWH F S WD

o rezolvare
FRQGXFHUHD

SRUQLQG

*HVWLXQHD

vQWUHSULQGHULL WUHEXLH DVWIHO I FXW VSXV JHVWLXQHD RSWLP WUHEXLH V

vQFkW V

VSRUHDVF

LD DF LRQDULORU VDX DOWIHO IXQF LHL GH XWLOLWDWH

XUP UHDVF

PD[LPL]DUHD

DF LRQDULORU 3XWHP IRUPXOD GHFL RSWLPXO PDQDJHULDO FD XQ PRGHO GH RSWLPL]DUH JHQHUDO 02* FX IXQF LD RELHFWLY

Max U(v1,v2,...,vT)

VXE

UHVWULF LL

WHKQLFH

HFRQRPLFH L ILQDQFLDUH FRUHVSXQ] WRDUH 5H]ROYDUHD XQXL DVWIHO GH PRGHO HYLGHQW GXS VSHFLILFDUHD UHVWULF LLORU FRQGXFH OD  SURJUDPXO GH LQYHVWL LL SH RUL]RQWXO >7@  SROLWLFD GH vPSUXPXW L GH SODVDPHQWH

3. politica distribuirii dividendelor pe orizontul ales.

5H]ROYDUHD

SUREOHPHL

SXVH

DQWHULRU

FRQVW

vQ

VHSDUDUHD

GHFL]LLORU



de

GHFL]LLOH  L  (D SRDWH IL I FXW GXS VFKHPD D PDQDJHULL DGRSW

SULQ UH]ROYDUHD 02* vQ PDQLHU

GHVFHQWUDOL]DW 

GHFL]LD  VWDELOLQG SURJUDPXO GH LQYHVWL LL SH ED]D XQRU

FULWHULL RELHFWLYH SXU HFRQRPLFH L WHKQLFH (L VXQW VXERUGRQD L DF LRQD

rilor, au rolul

GH D DOHJH SURLHFWHOH GH GH]YROWDUH L GH D SURSXQH DF LRQDULORU PRGXO GH ILQDQ DUH L

politica de distribuire a dividendelor;

E DF LRQDULL FX VDUFLQD DGRSW ULL GHFL]LLORU  L 

PRG

SUDFWLF

WUHEXLH

J VLWH

GRX

PRGHOH

GH

RSWLPL]DUH

00

0$

DVRFLDWH PDQDJHULORU L UHVSHFWLY DF LRQDULORU DVWIHO vQFkW

PRGHOXO

00

HVWH

LQGHSHQGHQW

GH

FRPSRUWDPHQWXO

DF LRQDULORU

GDW

GH

IXQF LD GH XWLOLWDWH 8

- VROX LD PRGHOXOXL 00 UHSUH]LQW -

R GDW

GH LQWUDUH vQ PRGHOXO 0$

VROX LD PRGHOXOXL JOREDO 02* VH RE LQH UH]ROYkQG PDL vQWkL 00 DSRL 0$ DVWIHO vQFkW GHOHJDUHD GH SXWHUH GH GHFL]LH FRQVLP LW QX VH WUDGXF SULQWU GH DF LRQDUL V

-o modificare a optimului managerial al ntreprinderii.

5.7. Specificarea modelului dinamic

Pentru a speFLILFD
FRPSOHW PRGHOXO SH OkQJ LSRWH]HOH GHMD I FXWH WUHEXLH

introduse unele suplimentare:

Ipoteza 7 (I7 ([FHGHQWXO EUXW GH H[SORDWDUH VDX PDUMD EUXW t HVWH R IXQF LH FUHVF 1: MBt= g(ANt-1)
QDWXU UDQGDPHQWH DYHUHD $FWLYXO WRDUH L VWULFW FRQFDY GH >@ FX FRQGL LLOH L H[SULP

UHDOL]DW

vQ SHULRDGD

nivelul activului net contabil la data t$FHDVWD HVWH R LSRWH] GH

g()0, g()<0
D

SXU HFRQRPLF

IDSWXO F

vQWUHSULQGHUHD VH J VHWH vQ VLWXD LD GH

GHVFUHVF WRDUH

1LYHOXO R

ctivului net contabil poate fi asimilat cu

SRWHQ LDO VWUDWHJLLORU GH GH D SURGXFH ERJ SH LH FDUH GH]YROWDUH

vQWUHSULQGHULL QHW FRQWDELO

ILLQG H[SULP

FDSDFLWDWH UH]XOWDWXO

ntreprinderea le-D
F PDUMD EUXW UHQWDELOLWDWHD

DGRSWDW SkQ

vQ DFHO PRPHQW 5HOD LD >@ DUDW  SH GH R SDUWH FH DFWLYXO QHW FUHWH  P VXU FH

FUHWH SH P VXU PDUJLQDO

g0 LDU SH GH DOW g<0

SDUWH F GDWRULW

VFDGH SH

DFWLYXO QHW FUHWH 

HIHFWXOXL FRQFXUHQ HL LVDX DO VDWXU ULL FHUHULL HWF

Ipoteza 8 (I8

GHILQHWH VWDUHD LQL LDO

D vQWUHSULQGHULL /D GDWD  DFWLYXO QHW

contabil este cunoscut, iar ntreprinderea nu are datorii: AN00 , X0=0.

Ipoteza 9 (I9 GHILQHWH VWDUHD ILQDO avea datorii: XT=0. (I10

LQYHVWL LLOH

 'H H[HPSOX OD GDWD

T, ntreprinderea nu va

Ipoteza 10

UHDOL]DWH

GH

vQWUHSULQGHUH

VXQW

LUHYHUVLELOH

It0,

t=1,2,...,T ,SRWH]D YL]HD]

H[FOXGHUHD GLQ DQDOL]

D SUREOHPHORU GH GH]LQYHVWLUH

Ipoteza 11 (I11
F

UDWD GREkQ]LL HVWH DFHHDL SHQWUX ILHFDUH SHULRDG 

rt=r

5H]XOW

FKHOWXLHOLOH ILQDQFLDUH VXQW GDWH GH UHOD LD

CFt= r.Xt-1. ANt-1 DGLF  At = a.ANt-1, unde a este

D FDSLWDOXOXL

Ipoteza 12 (I12
$FHDVWD vQVHDPQ

SROLWLFD GH DPRUWL]DUH D FDSLWDOXOXL vQWUHSULQGHULL HVWH OLQLDU  F

At HVWH SURSRU LRQDO

FX

rata de amortizare.

Ipoteza 13 (I13

HVWH GH IDSW R GHILQL LH D XQXL LQGLFDWRU GH HILFLHQ SULQ UHODWLD

vQWUHSULQGHULL 'HILQLP UDWD GH UHQWDELOLWDWH PDUJLQDO

(AN) = g (AN) - 1.
$FHDVW QR LXQH YD IL HVHQ LDO vQ DQDOL]D PRGHOXOXL (D H[SULP UDQGDPHQWXO

unui leu investit n ntreprindere, atunci cnd aceasta a investit deja AN lei n trecut (exclusiv amortizarea). Putem scrie acum modelul general al ntreprinderii (MOG): Max U(v1,v2,...vT) It + vt = xt + (1-)[ g(ANt-1) - rXt-1] + aANt-1 ANt - ANt-1 = It - aANt-1 Xt - Xt-1 = xt It 0 X0=XT=0 , AN0 = fixat , t=1,2,...,T

Vom trata acest model dinamic cu tehnicile de control optimal prezentate

DQWHULRU 3HQWUX vQFHSXW VWXGLHP FD]XO VLPSOLILFDW vQ FDUH OLPLW P RUL]RQWXO OD 

perioade (T=2 L QHJOLM

impozitele (W

 L DPRUWL]DUHD 

a=0).

0RGHO GH ED]

FX GRX

SHULRDGH

n ipotezele prezentate (T=2,W=0,a=0 mult: PDUMD EUXW UHDOL]DW vQ SHULRDGD  HVWH

PRGHOXO 02* VH VLPSOLILF

IRDUWH

MB1= g(AN0).
(D UHSUH]LQW SHULRDG  FDSDFLWDWHD GH DXWRILQDQ DUH GH FDUH GLVSXQH vQWUHSULQGHUHD SH DFHDVW

PDUMD EUXW

UHDOL]DW

vQ SHULRDGD  HVWH

MB2 = g(AN1) = g(AN0+I1),

ILLQG GHFL R IXQF LH GH LQYHVWL LLOH UHOL]DWH vQ SULPD SHULRDG 

VXPD vPSUXPXWDW

VDX SODVDW  vQ SHULRDGD  HVWH UDPEXUVDW LDU GDWRULW LSRWH]HORU ,

UHFXSHUDW L ,

) integral

n perioada 2. Deci, X2 = x1+x2+X0 CF1=0, CF2=rx1. - UHOD LLOH

XWLOL]DUH UHVXUVH

9,

avem: x2=-x1,

UHODWLYH OD FHOH GRX

SHULRDGH GHYLQ

I1+v1 = x1+MB1, respectiv, I2+v2 = g(AN0+I1) - (1+r)x1. Modelul (MOG) este n acest caz: Max U(v1,v2) I1+v1=x1+MB1 I2+v2=g(AN0+I1)-(1+r)x1 I10, I20.

Este vorba despre un model cu cinci variabile (I1,I2,x1,v1,v2

GRXD HJDOLW L L GRXD LQHJDOLW

L SDWUX UHVWULF LL

L 0RGHOXO GHSLQGH GH GRL SDUDPHWUL HVHQ LDOL

- nivelul activului net contabil AN0 FDUH SRDWH IL DVLPLODW P - rata dobnzii r.
3ROLWLFD RSWLPDO YD GHSLQGH GH DFHWL GRL

ULPLL vQWUHSULQGHULL

SDUDPHWUL

6H

YD

DQDOL]D

vQ

FRQWLQXDUH QXPDL FD]XO FkQG vQWUHSULQGHUHD HVWH GH WDOLH PLF V ILH LQFLWDW V LQYHVWHDVF  'XS FXP VH YD DU WD vQ

OD GDWD  DVWIHO vQ FRQWLQXDUH DFHVW

ct ea

FD] VH

SURGXFH DWXQFL FkQG HVWH vQGHSOLQLW &RQGL LD PDUJLQDO VH FkWLJ GH LQFLWDUH OD

FRQGL LD

(AN0) > r, sau AN0 < [g] (1+r).

-1
IDSWXO F UDWD GH UHQWDELOLWDWH UDWHL GREkQ]LL $OWIHO VSXV

LQYHVWLUH

VHPQLILF

D vQWUHSULQGHULL OD GDWD  HVWH VWULFW VXSHULRDU

PDL PXOW LQYHVWLQG vQ vQWUHSULQGHUH GHFkW SODVkQG EDQLL SH SLD D ILQDQFLDU 

&RQGL LD HVWH vQGHSOLQLW 3UHVXSXQHP F

pentru valori mici ale lui AN0 GDWRULW

IXQF LD GH XWLOLWDWH

FRQFDYLW

LL IXQF LH

i g.

HVWH GHULYDELO

L VWULFW FRQFDY  DGLF  DGLF QH VLWX P

U/v1>0, U/v2>0 'LIHUHQ LHP U(v1,v2) L SXQHP FRQGL LD dU=0

SH R FXUE GH L]RXWLOLWDWH

dU = (U/v1).dv1 + (U/v2).dv2 = 0.
5H]XOW 

U dv 2 v 1 = U dv 1 v 2

DP RE LQXW UDWD PDUJLQDO

GH VXEVWLWXLUH vQWUH OHX YHQLW GLQ SHULRDGD  L  OHX YHQLW

din perioada 1. Definim:

( v 1, v 2 ) = U U v 1 v 2 1

QXPLW

UDW

SVLKRORJLF

D DF LRQDULORU FDUH GHSLQGH GH YHQLWXUL L DUDW

F  OD

OLPLW  DF LRQDULL VXQW LQGLIHUHQ L vQWUH D GLVSXQH GH  OHX vQ SOXV vQ SHULRDGD  VDX

(1+ ) lei n perioada 2.

9RP VWDELOL FRQGL LLOH GH RSWLPDOLWDWH L YRP DQDOL]D RSWLPXO PDQDJHULDO DO vQWUHSULQGHULL SH QHOLQLDU  PRGHOXO FX GRX SHULRDGH FX DMXWRUXO WHKQLFLORU GH SURJUDPDUH

prezentate n capitolul 4.

/DJUDQJHDQXO DWDDW PRGHOXOXL HVWH

L = U(v1, v2) + 1I1 + 2I2 - 1[ I1+v1-x1-MB1 ]- 2[ I2+v2-g(AN0+I1) +(1+r)x1].

6H YHULILF LPHGLDW F F /

() HVWH R IXQF LH VWULFW FRQFDY

-Tucker:

vQ YDULDELOHOH

v1, v2, x1,

I1, I2 FHHD FH vQVHDPQ

UHOD LLOH .XKQ 7XFNHU VXQW FRQGL LL QHFHVDUH L VXILFLHQWH GH

RSWLPDOLWDWH 6FULHP FRQGL LLOH .XKQ

 $QXO P GHULYDWHOH SDU LDOH DOH OXL /

() n raport cu variabilele v1, v2, x1, I1, I2:

U/v1 -J1=0

L/v1 sau

 LPSOLF

1=U/v1
L/v2
 LPSOLF

[1] U/v2-J2=0 [2]

sau 2=U/v2

L/x1=0 LPSOLF 1-2(1+r)=0 [3] L/I1=0 LPSOLF 1 -J1+J2g(AN0+I1)=0 [4] L/I2=0 LPSOLF 2 -J2 =0
[5].

 0XOWLSOLFDWRULL DWDD L UHVWULF LLORU GH LQHJDOLWDWH VXQW QHQHJDWLYL

1 0
1I1


>@

2 0 [7]
2I2=0 [9].

 &RQGL LLOH GH WUDQVYHUVDOLWDWH VXQW >@ L

&HOH  UHOD LL vPSUHXQ

FX UHVWULF LLOH PRGHOXOXL SHUPLW GHWHUPLQDUHD FRPSOHW D VROX LHL RSWLPH

* * ) DGLF a variabilelor ( v 1* , v 2 , x 1* , I 1* , I 2

3URSULHW

L DOH VROX LHL RSWLPH

Proprietatea 1 'DF
SULPD SHULRDG  DGLF

HVWH vQGHSOLQLW

FRQGL LD GH LQFLWDUH vQWUHSULQGHUHD LQYHVWHWH vQ

I*1>0.
J

'HPRQVWUD LH 'LQ UHOD LLOH >@ L >@ VH HOLPLQ

L RE LQHP

1+J2[g(AN0+I1)-(1+r)]=0
Deoarece U(v1,v2)
RE LQHP HVWH VWULFW FRQFDY  GLQ >@ UH]XOW J

[10]
2>0
PS U LP >@ OD J

1/J2 + [g(AN0+I1)-(1+r)] = 0
&RQGL LD GH LQFLWDUH OD LQYHVWLUH VWDELOLW SULQ DEVXUG F FDQWLWDWH DQWHULRU HVWH

[11]. g(AN0)-(1+r)>0. Presupunem,

HVWH R DU WUHEXL FD

I1=0

$WXQFL vQ UHOD LD >@ H[SUHVLD GLQ SDUDQWH]D GUHDSW FD UHOD LD >@ V ILH DGHY UDW 

SR]LWLY  J

3HQWUX

1/J2<0.

Deoarece
F

2>0,

rezXOW 1<0


FHHD FH FRQWUD]LFH FRQGL LD GH RSWLP >@ 3UHVXSXQHUHD

I1=0 HVWH GHFL IDOV

Proprietatea 2 QWUHSULQGHUHD QX LQYHVWHWH vQ SHULRDGD  DGLF I*2=0.

'HPRQVWUD LH 5HOD LD >@ DUDW >@ G F

2 = 2 $P Y
LQYHVWHWH vQ

]XW F

2>0 GHFL L 2 > 0 5HOD LD

SHULRDG DVWIHO vQFkW UDWD

atunci, n mod necesar, I2 = 0.

QWUHSULQGHUHD ILH HJDO SULPD LL PDUJLQDOH V FX UDWD GREkQ]LL DGLF  F GDF

Proprietatea 3
UHQWDELOLW

(AN0+I*1) = r.
5HOD LD >@

'HPRQVWUD LH 'LQ UHOD LD >@ UH]XOW

I*1>0, trebuie ca 1=0

L FXP J

devine n acest caz:

2[g(AN0+I1)-(1+r)]=0
DGLF

2>0,

avem g(AN0+I1) = 1+r,

PDUJLQDOH GH

sau g(AN0+I1) - 1 = r

(AN0+I*1)=r FRQIRUP FX GHILQL LD UDWHL

rentabilitate a ntreprinderii. Proprietatea 4. 9HQLWXULOH DF LRQDULORU VXQW IL[DWH vQ ILHFDUH SHULRDG

SVLKRORJLF D DF LRQDULORU V ILH HJDO FX UDWD GREkQ]LL DGLF  DVWIHO vQFkW UDWD

* v 1* , v 2 = r

'HPRQVWUD LH 'LQ UHOD LD >@ UH]XOW UHOD LLOH >@ L >@ DYHP 3URSULHW

1+r=1/2, sau r=(1/2)-1

F

LQkQG FRQW GH

* . r = v 1* , v 2
RSWLPXO PDQDJHULDO DO vQWUHSULQGHULL QRWDW

LOH GHPRQVWUDWH DUDW

* * prin ( ) VH GHWHUPLQ v 1* , v 2 , x 1* , I 1* , I 2

GLQ UHOD LLOH

(AN0+I*1) = r
* v 1* , v 2 = r

[12]

I*1+v*1 = x*1+g(AN0) v*2=g(AN0+I*1)-(1+r)x*1

$P Y ]XW F

1=0 L UHOD LD >@ GHYLQH g(AN0+I*1)-(1+r)=0

L VXILFLHQW

SXWkQG IL VFULV IRUP 

sub forma g(AN0+I*1)/(1+r)=1 sau -1+[g(AN0+I*1)]/(1+r)=0

UHOD LD QX HVWH DOWFHYD GHFkW FRQGL LD QHFHVDU

6XE DFHDVW

GH RSWLPDOLWDWH

pentru

problema: Max Z = - I1 + [g(AN0+I1)]/(1+r) ntr-DGHY

)XQF LD U GLQ FRQGL LD HVWH

[13]

Z/I1

 UH]XOW  D

- 1+[ g(AN0+I1)]/(1+r)=0.

YDORDUHD

DFWXDOL]DW

LUXOXL

{-I1, g(AN0+I1)} de fluxuri

OD PDL

ILQDQFLDUH QHWH GDWRUDWH LQYHVWL LLORU O PXOWH SHULRDGH L FRQGXFH OD WHRUHPD

a nivelul I1

5H]XOWDWXO VH JHQHUDOL]HD]

7HRUHP  Q LSRWH]HOH PRGHOXOXL DEVHQ D LPSR]LW ULL SURILWXOXL L SLD SHUIHFW  RSWLPXO PDQDJHULDO DO vQWUHSULQGHULL FRUHVSXQGH FX

GH FDSLWDO

PD[LPL]DUHD YDORULL

acWXDOL]DWH D LUXOXL GH IOX[XUL ILQDQFLDUH JHQHUDW GH LQYHVWL LLOH HIHFWXDWH

Putem utiliza, deci, drept criteriu de optimizare, maximizarea valorii actualizate nete (VAN).
([DPLQkQG UHOD LLOH >@ REVHUY P F QLYHOXO LQYHVWL LLORU HVWH GHWHUPLQDW GH

priPD

UHOD LH DGLF

LQGHSHQGHQW GH FHOHODOWH YDULDELOH GH GHFL]LH 3XWHP GHFL IL[D V QH SUHRFXSH SUREOHPHOH OHJDWH GH GLVWULEXLUHD GLYLGHQGHORU

QLYHOXO LQYHVWL LLORU I U

VDX SROLWLFD GH vPSUXPXW $FHDVWD vQVHDPQ FXP DP DU WDW DQWHULRU 0DQDJHULL vQWUHSULQGHULL FRQWDELOL]HD] OHJDWH GH LQYHVWL LL L GHWHUPLQ 

GHFL]LD SRDWH IL GHVFHQWUDOL]DW  DD

YHQLWXULOH QHWH

~ v1

~ v 2 care sunt direct

~ v 2 = MB2-I2 = g(AN0+I1)-I2. v 1 =MB1- I1 L ~

Putem scrie acum modelul managerului (MM) astfel: Max [ ~ v1 + ~ v 2 /(1+r) ]

~ v 1 MB1
~ v 2 g[AN0+(MB1- ~ v 1 )]

Modelul este echivalent cu [13], la care s-DX

LUHYHUVLELOLWDWH GHWHUPLQ LQYHVWL LL D LQYHVWL LLORU vQ FHOH GRX SHULRDGH SROLWLFD RSWLPDO

DG XJDW

UHVWULF LLOH 00

GH L GH

0DQDJHUXO

UH]ROY

(~ v ,~ v )
* 1

* 2  GH XQGH YD UH]XOWD LPHGLDW SROLWLFD RSWLP

(I*1,I*2):
* I*1 = MB1- ~ v1 * I*2 = g(AN0+I*1)- ~ v2 .

6H

YHULILF

XRU

DFHDVWD

HVWH

VROX LH

RSWLP

VFULLQG

ODJUDQJHDQXO

DWDDW

(MM): L =~ v 1 )+2[g(AN0+(MB1- ~ v 1 ))- ~ v1 + ~ v 2 /(1+r)+1(MB1- ~ v2 ]

&RQGL LLOH .XKQ 7XFNHU VXQW XUP WRDUHOH

1. L/ ~ v1 =0 LPSOLF 1-O1-O2g(AN0+(MB1- ~ v 1 ))=0

L/ ~ v 2 =0 LPSOLF 1/(1+r)-O2=0
2. 10 , 20 3. 1(MB1- ~ v 1 )=0, 2[g(AN0+(MB1- ~ v 1 ))- ~ v 2 ]=0 .
6ROX LD WUDQVPLV RSWLP

(~ v ,~ v )
* 1 * 2

GHILQHWH

SROLWLFD

RSWLP

PDQDJHULORU

(D

HVWH

DF LRQDULORU FRQVWLWXLQG R GDW

GH LQWUDUH vQ

modelul acestora (MA). n acest

fel, politica

GH

LQYHVWL LL

vQWUHSULQGHULL

QX

GHSLQGH

GH

SUHIHULQ HOH

DF LRQDULORU

UHSUH]HQWDWH SULQ IXQF LD GH XWLOLWDWH

U). Altfel spus,

LQYHVWL LLOH VH DOHJ SH ED]D

criteriilor

pur economice (r L WHKQLFH g L QX SH ED]D SUHIHULQ HORU DF LRQDULORU U).

0RGHOXO DF LRQDULORU 0$ HVWH

Max U(v1, v2)

* v1= ~ +x1 v1 * v2 = ~ -(1+r)x1 v2

sau, eliminnd pe x1 vQWUH FHOH GRX

UHOD LL

Max U(v1, v2)

* * v1 +~ v2 /(1+r). v1 + v2/(1+r) = ~

,QWHUSUHWDUHD PRGHOXOXL 0$ HVWH LPHGLDW  SROLWLFD RSWLPDO FRQVW vQ PD[LPL]DUHD IXQF LHL ILH HJDO GH XWLOLWDWH D YHQLWXULORU VXE

D DF LRQDULORU FD VXPD

UHVWULF LD  HJDO

YHQLWXULORU DFWXDOL]DWH V

FX R FRQVWDQW

* * [~ v1 +~ v2 /(1+r)] DGLF
D PDQDJHULORU

FX VXPD

YHQLWXULORU DFWXDOL]DWH UH]XOWDW

GLQ SROLWLFD RSWLP

6ROX LH JUDILF

$QDOL]D VROX LHL PRGHOXOXL HVWH SUH]HQWDW

vQ ILJXUD QU 

n planul (v1, v2 v0=g(AN2+MB1-v1)

$FHDVW PDQDJHULORU (D HVWH R FXUE

VH

WUDVHD] 

SHQWUX

v1(-, MB1) G
UHSUH]LQW

FXUED

GH

HFXD LH

FXUE  QRWDW

SH JUDILF 

FXUED YHQLWXULORU

GHVFUHVF WRDUH L FRQFDY  &XUED WDLH D[D

Ov2 n punctul

A(0,(AN0+g(AN0)))

L VH RSUHWH vQ SXQFWXO

B( g(AN0), g(AN0)).

v2 N* (U) A (G) M*

B I*1 () v1 D

)LJXUD QU  6ROX LD PRGHOXOXL 0$

0XO LPHD SXQFWHORU VLWXDWH VXE FXUED 

G) corespunde veniturilor managerilor

SHULRDGH VXQW SR]LWLYH VDX vQ SXQFWXO

* ~* (~ , v2  v1

SHQWUX FDUH FKHOWXLHOLOH GH LQYHVWL LL vQ FHOH GRX D PDQDJHULORU HVWH UHSUH]HQWDW

QXOH 3ROLWLFD RSWLP GUHDSWD GH HFXD LH

* ~* M*( ~ , v 2 ), n care v1

* * v1+v2/(1+r)= ~ +~ /(1+r), v1 v2

notaW

SH JUDILF HVWH WDQJHQW

OD FXUED 

G). Toate punctele situate pe dreapta ()

DFWXDOL]DW 

FRUHVSXQG YHQLWXULORU FDUH GDX DFHHDL YDORDUH QHW

* * VAN*= ~ +~ /(1+r). v1 v2
HVWH HJDO

n particular, abscisa punctului D

HVWH QXO 

LQWHUVHF LD

dreptei () cu axa Ov1

tocmai cu VAN*. Punctul M* fiind situat pe curba (G LQYHVWL LD RSWLP I*2=0

,QYHVWL LD RSWLP L D SULPHL SHULRDGH HVWH P VXUDW RSWLP D DF LRQDULORU HVWH

D SHULRDGHL  SULQ GLIHUHQ D

absciselor punctelor M*

B

3ROLWLFD

UHSUH]HQWDW HVWH WDQJHQW

GH OD

punctul N*(v*1,v*2) n care curba (U

GH HFXD LH

U(v1,v2)=U(v*1,v*2)

dreapta ( etape:

$VWIHO RSWLPXO PDQDJHULDO DO vQWUHSULQGHULL VH GHWHUPLQ

JUDILF vQ GRX

VH RSWLPL]HD] VH RSWLPL]HD]

VAN pe curba (G L VH GHILQHWH GUHDSWD );

IXQF LD GH XWLOLWDWH

U pe dreapta ().
L 

Pantele curbelor (G) n punctul M* dreptei (

$FHDVWD H[SULP HJDOLWDWHD DF LRQDULORU L UDWD UHQWDELOLW

U) n punctul N* sunt egale cu panta

UDWD GREkQ]LL UDWD SVLKRORJLF D

vQWUH

LL PDUJLQDOH D vQWUHSULQGHULL IDSWXO F SXQFWXO

&RQGL LD GH LQFLWDUH OD LQYHVWLUH LPSOLF

M*

VH VLWXHD]

OD

stnga punctului B $FHDVW

FRQGL LH VH SRDWH VFULH VXE IRUPD

-(1+r) -g(AN0),
DVWIHO F SDQWD GUHSWHL 

 HVWH VXSHULRDU

FHOHL D WDQJHQWH

i la curba (G) n punctul B. M*

VH VLWXHD] vQ

6H GLVWLQJ GRX

FD]XUL

D FkQG LQYHVWL LD HVWH vQ vQWUHJLPH DXWRILQDQ DW  SXQFWXO

primul cadran, ntre punctele A L B ( HVWH LQIHULRDU

WDQJHQWHL vQ SXQF

$VWIHO GH VLWXD LL DSDU DWXQFL FkQG SDQWD GUHSWHL

tul A la curba (G):

(1+r)g[AN0+g(AN0)].
,QHJDOLWDWHD DQWHULRDU P ULPLL LQL LDOH D VH YHULILF  SHQWUX YDORUL PDUL DOH UDWHL GREkQ]LL LVDX DOH vQWUHSULQGHULL 

AN0

$XWRILQDQ DUHD

HVWH

VXILFLHQW

SHQWUX

DFRSHULUHD FKHOWXLHOLORU GH LQYHVWL LL DWXQFL

cnd:
GHRDUHFH UDWD GREkQ]LL IDFH FD RSHUD LLOH

LQYHVWL LD RSWLPDO V

HVWH VODE

GH SODVDPHQW SH SLD D ILQDQFLDU

ILH PDL DWUDFWLYH GDWRULW P ULPLL VDOH GH FDSDFLWDWH GH

vQWUHSULQGHUHD

GLVSXQH

DXWRILQDQ DUH VXILFLHQW

SHQWUX D QX UHFXUJH OD DOWH VXUVH GH ILQDQ DUH

E FkQG LQYHVWL LLOH VXQW SDU LDO DXWRILQDQ DWH SXQFWXO DO GRLOHD DGLF OD VWkQJD DGLF SXQFWXOXL

M* este situat n cadranul

DSDU DWXQFL FkQG D P ULPHD LQL LDO

A

$VWIHO

GH

VLWXD LL

(1+r)<g[AN0+g(AN0)]

DWXQFL FkQG

UDWD GREkQ]LL LVDX

vQWUHSULQGHULL LDX YDORUL PLFL Q DVWIHO GH VLWXD LL LQYHVWL LLOH VXQW ILQDQ DWH SH GH R

SDUWH

SULQ

DXWRILQDQ DUH

LDU

SH

GH

DOW

SDUWH

SULQ

vPSUXPXW

LVDX

FUHWHUHD

capitalului.

([WHQVLL DOH PRGHOXOXL GH ED]

FX GRX

SHULRDGH

Modelul pre]HQWDW

L DQDOL]DW DQWHULRU SRDWH IL H[WLQV SULQ DG XJDUHD GH QRL

LSRWH]H $VWIHO VH SRDWH YHGHD FDUH HVWH LPSDFWXO OX ULL vQ FRQVLGHUDUH D LPSR]LWXOXL SH SURILW L DPRUWL] ULL VDX D XQRU FRQGL LL GH vPSUXPXW GLIHULWH SHQWUX PDQDJHUL SHUVRDQH MXULGLFH L DF LRQDUL SHUVRDQH IL]LFH ,DW  GH H[HPSOX FH GHYLQH PRGHOXO vQ FD]XO OX ULL vQ FRQVLGHUDUH D LPSR]LWXOXL L DPRUWL] ULL

Max U(v1,v2) I1+v1 = x1+(1-)MB1 I2+v2=(1-)[g((1-a)AN0+I1)-rx1]-x1+ a((1-a)AN0+I1) I1 0, I2 0.

,QWRGXFHP QRWD LLOH

r = (1-)r - UDWD UHDO

D GREkQ]LL GXS LPSR]LWDUH vQ SHULRDGD  U DPRUWL]DUH QHW  GXS LPSR]LWDUH

MB1 = (1-)MB1 -PDUMD QHW

AN0 = (1-a)AN0 -DFWLYXO QHW FRQWDELO I

g ( AN0 +I1)=(1-)[g((1-a)AN0+I1)+a((1-a)AN0+I1) -PDUMD

UHDOL]DW vQ SHULRDGD 

( AN0 +I1)=(1-)((1-a)AN0+I1) -rata de rentabilitate

&RQGL LD GH LQFLWDUH OD LQYHVWLUH GHYL

PDUJLQDO

GXS

LPSR]LWDUH

ne:

((1-a)AN0) > r
&X DFHVWH QRWD LL PRGHOXO GHYLQH

Max U(v1,v2) I1+v1 = x1+ MB1 I2+v2 = g ( AN0 +I1)-(1+ r )x1

I1 0, I2 0
DGLF H[DFW PRGHOXO

precedent, n care s-a nlocuit rata dobnzii (r) cu rata dobnzii



reale ( r  L PDUMD EUXW

MB1 FX PDUMD QHW

MB1 ).

6H YRU RE LQH GHFL UH]XOWDWH DQDORDJH

- optimul managerial este caracterizat de egalitatea diQWUH

DF LRQDULORU UDWD GH UHQWDELOLWDWH PDUJLQDO LQYHVWL LLOH RSWLPDOH VXQW VHOHFWDWH GXS

UDWD

SVLKRORJLF

L UDWD GREkQ]LL UHDOH

FULWHULXO 9$1 FX UDWD GH DFWXDOL]DUH HJDO

cu rata dobnzii reale; SULQFLSLXO VHSDU ULL GHFL]LLORU vQWUH DF LRQDUL L PDQDJHUL U PkQH YDODELO

Aplicarea tehnicilor de control optimal

Am tratat anterior modelul de dezvoltare a ntreprinderii prin prisma tehnicilor

GH SURJUDPDUH QHOLQLDU  9RP DERUGD YRU vQ FRQWLQXDUH GRX DFHODL PRGHO XWLOL]kQG vQV VH 6H VWXGLD FD]XUL FkQG vQWUHSULQGHUHD

tehnicile de control optimal.

DXWRILQDQ HD] L FkQG DSHOHD] U

OD vPSUXPXW GH SH SLD D ILQDQFLDU 

D 0RGHOXO I

vPSUXPXW

Max VAN = v1+v2/(1+r) I1+v1 = g(AN0) I2+v2 = g(AN1) AN1-AN0 = I1-aAN0 I1 g(AN0) I2 g(AN1) I1 0, I2 0
3HQWUX D WUHFH OD XQ PRGHO GLQDPLF FRQWLQXX SUHVXSHXQHP F RUL]RQWXO GH WLPS

este infinit [0, 5HOX

DFFHGH OLEHU OD SLD D

P QRWD LLOH LQWURGXVH DGDSWkQGX

-le la timpul continuu. Pentru arii pot

GH UDWD FDUDFWHUL]DW

VLPSOLILFDUH YRP FRQVLGHUD F

vQWUHSULQGHUHD QX SO WHWH LPSR]LW LDU DF LRQ SLD SUHVXSXV SHUIHFW 

FDSLWDOXULORU

dobnzii r FRQVWDQW

SH WRW RUL]RQWXO GH WLPS

 fie AN(t) nivelul activului net contabil la data t. La data 0, acesta este AN(0)=AN0; ntre momentele t L t+dt, QWUHSULQGHUHD UHDOL]HD]
HIHFWXHD] FKHOWXLHOLOH GH D LQYHVWL LL R PDUM EUXW HJDO FX

MB(t)dt,

I(t)dt, distribuie dividendele [MB(t)-

I(t)]dt L DPRUWL]HD] la data t LDU DFHDVW

R SDUWH

GLQ FDSLWDOXO V X

PDUMD EUXW  SH XQLWDWHD GH WLPS

MBW HVWH R IXQF LH g de capitalul ntreprinderii

IXQF LH QX GHSLQGH GH WLPS   L J PDUMD 

MB(t)=g(AN(t)), FX FRQGL LLOH g  J

FKHOWXLHOLOH GH LQYHVWL LL QX SRW IL PDL PDUL DXWRILQDQ DUH

;
UHVWULF LD GH

GHFkW

EUXW

I(t) g(AN(t)); versibile: I(t) 0.

LQYHVWL LLOH VXQW LUH

Sub forma unui program de control optimal, modelul se scrie astfel: Max [g(AN(t) - I(t)]e - rt dt
0

AN (t) = I(t)-aAN(t)

g(AN(t))-I(t)0 I(t)0 AN(0)=AN0. Este vorba despre un program de control optimal care cRPSRUW stare, AN R YDULDELO orizont infinit.
&RQGL LLOH QHFHVDUH GH RSWLPDOLWDWH VXQW GDWH GH SULQFLSLXO PD[LPXOXL DO OXL FRQVLGHUkQG F UHVWULF LLOH PRGHOXOXL GH FRQWURO R YDULDELO GH

I UHVWULF LL DVXSUD YDULDELOHORU GH VWDUH L GH FRQWURO

Pontreaghin. Definim hamiltonianul problemei,

VH SRW WUDWD FD vQ ODJUDQJHDQXO FODVLF L UHQXQ kQG V

PDL VSHFLILF P GH ILHFDUH GDW

variabila t: H=g(AN)-I +(I - aAN) +I + (g(AN) - I)

HVWH R IXQF LH FRQFDY vQ L

AN

I

GHFL FRQGL LLOH QHFHVDUH GH RSWLPDOLWDWH YRU IL L

VXILFLHQWH $FHVWH FRQGL LL VXQW

 =(a+r)-(1+)g(AN)

[1] [2] [3] [4]

-(1+) +=0 I=0, 0, (g(AN)-I)=0, 0

t

lim (t)e -rt = 0

5HOD LD >@ HVWH FRQGL LD GH WUDQVYHUVDOLWDWH D SUREOHPHL SHQWUX RUL]RQW LQILQLW L H[SULP IDSWXO F LQWHJUDOD GLQ FULWHULXO GH RSWLP HVWH ELQH GHILQLW  L

Multiplicatorii
FRQVWLWXLW GLQ

definesc patru regimuri, dintre care numai trei sunt

YD IL DFHVWH SH GXUDWD SULQ F URUD vQWUHSULQGHUHD DGRSW XQXO GLQ

realizabile SHQWUX F L QX SRW IL VLPXOWDQ VWULFW SR]LWLYL 7UDLHFWRULD RSWLPDO

SHULRDGH UHJLPXUL (OH VH FDUDFWHUL]HD]

- regimul 1 (>0,
5HOD LD YHULILF 0 ULPHD >@

 QWUHSULQGHUHD QX LQYHVWHWH QLPLF

ci distribuie dividendele.
F YDULDELOD DG M X Q FW

GHYLQH

-1+=0

FHHD

FH

LPSOLF

IDSWXO

LQHJDOLWDWHD vQWUHSULQGHULL

(t)<1 L VDWLVIDFH HFXD LD GLIHUHQ LDO =(a+r)-g(AN).

GHVFUHWH FX UDWD

a, deoarece AN (t)=-aAN(t). De

DVHPHQHD PDUMD EUXW

L GLYLGHQGHOH FDUH VXQW HJDOH vQ DFHVW FD] GHVFUHVF

- regimul 2 (=0,

 QWUHSULQGHUHD LQYHVWHWH R SDUWH GLQ PDUM  LDU UHVWXO R F YDULDELOD DGMXQFW HVWH R FRQVWDQW

GLVWULEXLH FD GLYLGHQGH 5HOD LD >@ DUDW

(t)=1,

LDU HFXD LD GLIHUHQ LDO F

>@ FRQGXFH OD

g(AN)=a+r, sau g(AN)-a=r, ceea

FX UDWD

FH vQVHDPQ

UHQWDELOLWDWHD PDUJLQDO

PLQXV DPRUWL]DUHD HVWH HJDO

GREkQ]LL QWUHSULQGHUHD PHQ LQH P ULPHD VD OD XQ QLYHO FRQVWDQW UHSUH]LQW VROX LD HFXD LHL

AN*, nivel ce

g(AN)=a+r

5H]XOW

vQWUHSULQGHUHD LQYHVWHWH GRDU

pentru a compensa amortizarea: I = aAN*. - regimul 3 (=0,! QWUHSULQGHUHD LQYHVWHWH WRDW

P ULPHD 9DULDELOD vQWUHSULQGHULL DGMXQFW HYROXHD] GXS HFXD LD PDUMD

I(t)=g(AN(t)) DVWIHO F

GLIHUHQ LDO L OXL

AN (t)=g(AN(t))-aAN(t).
HFXD LD UHOD LD GLIHUHQ LDO >@ L GLQ

YHULILF

LQHJDOLWDWHD SULQ

(t)>1

VDWLVIDFH 

=[(a+r)-g(AN)] 

RE LQXW

HOLPLQDUHD

GLQ

hamiltonian.

5HJLPXULOH  L  SRW IL FDUDFWHUL]DWH IXQF LHL J UHQWDELOLWDWHD PHGLH

LQkQG FRQW F HVWH PDL

VXE LSRWH]HOH I FXWH DVXSUD PDUH GHFkW UHQWDELOLWDWHD

LQVWDQWDQHH

PDUJLQDO  DGLF 

g(AN)/AN>g(AN), pentru orice AN0.

3URSR]L LD 

Atunci cnd ntreprLQGHUHD

DGRSW

UHJLPXO  GLYLGHQGHOH Y UVDWH VXQW

strict pozitive.
'HPRQVWUD LH

Deoarece aa+r, atunci a+r=g(AN*)

DGLF

g(AN*)<g(AN*)/AN*,

UH]XOW

ag(AN*)/AN*
UHSUH]LQW 3URSR]L LD

aAN*<g(AN*), sau g(AN*)-aAN*>0

vQWUHSULQGHUHD DGRSW vQ

([SUHVLD GLQ SDUWHD VWkQJ

WRFPDL GLYLGHQGHOH Y UVDWH GH vQWUHSULQGHUH DF LRQDULORU  'DF SHUPDQHQ UHJLPXO  P ULPHD

ntreprinderii tinde spre o valoare AN mai mare dect AN*.

'HPRQVWUD LH (FXD LD GLIHUHQ LDO

AN (t)=g(AN(t))-aAN(t)

admite un punct fix AN 

IXQF LD UH]XOW J
HVWH

FDUH HVWH VROX LD HFXD LHL

g(AN)-aAN = 0
QHJDWLY  GLQ

$YHP UHOD LLOH

a=g( AN )/AN, g( AN )<g( AN )/ AN =a<a+r=g(AN*), DGLF g( AN )<g(AN*). Deoarece

GHVFUHVF WRDUH GHULYDWD VD J HVWH

g(AN*)>g( AN )

AN >AN .
)LHFDUH GLQ FHOH WUHL UHJLPXUL FRUHVSXQG PXO LPLL GH GHFL]LL FDUDFWHULVW

ice unei

strategii a ntreprinderii. Astfel, regimul 1 corespunde unei strategii de abandon.

QWUHSULQGHUHD vQFHWHD] V PDL LQYHVWHDVF  UHGXFkQGX L

SURJUHVLY

DFWLYLWDWHD

5HJLPXO  HVWH UHJLPXO SHUPDQHQW DO PRGHOXOXL QWUHSULQGHUHD PHQ LQH DFWLYLWDWHD XQ QLYHO FDUH FRUHVSXQGH RSWLPXOXL PDQDJHULDO SH WHUPHQ OXQJ (D LQYHVWHWH

la
L

GLVWULEXLH R SDUWH GLQ SURILWXUL DF LRQDULORU 5HJLPXO  HVWH XQ UHJLP GH FUHWHUH PD[LPDO  DVRFLDW XQHL VWUDWHJLL GH GH]YROWDUH QWUHSULQGHUHD FRQVDFU WRW SURILWXO

pentru LQYHVWL LL UHVSHFWkQG FRQGL LD GH DXWRILQDQ DUH

b) Modelul cu mprumut. Vom generaliza modelul precedent n cazul cnd ntreprinderea poate utiliza
SRVLELOLW LOH GH vPSUXPXW OD R UDW N PDL PDUH GHFkW UDWD GREkQ]LL 

k>r).

1RW P FX

X(t) suma datoriei ntreprinderii la data t L FX x(t) YDULD LD GDWRULHL vQ

3UHVXSXQHP F OD GDWD  vQWUHSULQGHUHD QX DUH GDWRULL

unitatea de timp la data t

$YHP UHOD LLOH

X (t)=x(t), X(0)=0, X(t) 0. Nivelul datoriei, n orice moment, nu

SURSRU LH T GLQ IRQGXULOH SURSULL DGLF 

poate dHS

L R DQXPLW

X(t)q[AN(t) - X(t)], sau, X(t) sAN(t), cu s=q/(1-q)<1.

Q DFHVWH FRQGL LL PRGHOXO VH VFULH VXE IRUPD XQXL SURJUDP GH FRQWURO RSWLPDO

astfel: Max [g(AN(t)) - I(t) - kX(t) + x(t)]e - rt dt

0

AN (t)=I(t)-aAN(t)
X (t)=x(t)

I(t)0; g(AN(t))-I(t)-kX(t)+x(t)0; X(t) sAN(t) X(t) 0, X(0)=0, AN(0)=AN0

0RGHOXO HVWH FDUDFWHUL]DW GH GRX YDULDELOH GH VWDUH  GH LQHJDOLWD

AN

X

GRX

YDULDELOH

de control (I hamiltonianul:

GH

SDWUX

UHVWLF LL

te legate ntre ele. Scriem

H=(1+)[g(AN)-I+x-rX]+(I-aAN)+x+I+X+(sAN-X).
&RQGL LLOH GH RSWLPDOLWDWH VXQW

= (a+r)-(1+)g(AN) = r+ k(1+)+-

-(1+ )+ =0
-(1+) +=0

I=0, 0, (g(AN)-I-kX+x)=0, 0 (sAN-X)= X=0, 0, 0.

$FHVWH FRQGL LL VXQW FDUDFWHUL]DWH GH SDWUX PXOWLSOLFDWRUL 

,,,) care definesc

 UHJLPXUL GLIHULWH GDU QX WRDWH VXQW UHDOL]DELOH / V P FD H[HUFL LX FDUDFWHUL]DUHD

acestor regimuri L DQDOL]D WUDLHFWRULHL RSWLPDOH

REZUMATUL CAPITOLULUI

1 /XDUHD vQ FRQVLGHUDUH H[SOLFLW

controlului optimal.
 8Q PRGHO GLQDPLF HVWH

D WLPSXOXL FRQGXFH OD PRGHOH GLQDPLFH FDUH VH UH]ROY

FX PHWRGD

FDUDFWHUL]DW

GH

XUP WRDUHOH

HOHPHQWH

RUL]RQWXO

VF

ara de timp,
L FULWHULXO GH

YDULDELOHOH GH VWDUH YDULDELOHOH GH FRQWURO HFXD LD GH WUDQ]L LH D VW ULORU VWDUHD LQL LDO

optimizare.
 'XS PRGXO GH FRQVLGHUDUH D YDULDELOHL WLPS PRGHOH GLQDPLFH SRW IL GLVFUHWH VDX FRQWLQXL FRQGL LLOH QHFHVDUH GH RSWLPDOLWDWH vQ PRGHOHOH GLQDPLFH

4. Principiul maximului fuUQL]HD]

 5HFRPDQG UL ELEOLRJUDILFH >@ >@ >@ >@ >@ D