• No results found

Strong Large Deviations Principles of Non-Freidlin-Wentzell Type - Optimal Control Problem with Imperfect Information - Jumps Phenomena in Financial Markets

N/A
N/A
Protected

Academic year: 2021

Share "Strong Large Deviations Principles of Non-Freidlin-Wentzell Type - Optimal Control Problem with Imperfect Information - Jumps Phenomena in Financial Markets"

Copied!
134
0
0

Loading.... (view fulltext now)

Full text

(1)

ISSN 2201-7372

Volume 2, Number 2, 2014, 230-363

Β© Copyright 2014 the authors. 230

Strong Large Deviations Principles of Non-Freidlin-Wentzell Type -Optimal Control Problem with Imperfect Information

-Jumps Phenomena in Financial Markets

J. Foukzon

Israel Institute of Technologies, Department of Mathematics, Haifa, Israel

Abstract. The paper presents, a new large deviations principles (SLDP) of non-Freidlin-Wentzell type, corresponding to the solutions Colombeau-Ito’s SDE. Using SLDP we present a new approach to construct the Bellman function 𝑣(𝑑, 𝒙) and optimal control 𝒖(𝑑, 𝒙) directly by way of using strong large deviations principle for the solutions Colombeau-Ito’s SDE. As important application such SLDP, the generic imperfect dynamic models of air-to-surface missiles are given in addition to the related simple guidance law. A four, examples have been illustrated proposed approach and corresponding numerical simulations have been illustrated and analyzed. Using SLDP approach, Jumps phenomena, in financial markets, also is considered. Jumps phenomena, in financial markets is explained from the first principles, without any reference to Poisson jump process. In contrast with a phenomenological approach we explain such jumps phenomena from the first principles, without any reference to Poisson jump process.

Keywords: Optimal control, Bellman equation, Colombeau-Ito’s SDE,Large deviations principles, Algebra of Colombeau generalized functions, Poisson jump process, Jumps phenomena, in financial markets.

(2)

1. Introduction

What new scalable mathematics is needed to replace the traditional Partial Differential Equations (PDE) approach to differential games?

Letβ„­ = (𝛺, 𝛴. 𝑃)be a probability space. Any stochastic process onℝ𝑛 is aΞ£-measurable

mapping𝑋: 𝛺 Γ— [0, 𝑇] β†’ ℝ𝑛. Many stochastic optimal control problems essentially come

down to constructing a function𝑒(𝑑, π‘₯) that has the properties: (1) 𝑒(𝑑, π‘₯) = inf𝛼�𝐉̅ ��𝑋𝑠,𝐷π‘₯ (πœ”)οΏ½π‘Žβˆˆ[0,𝑑]; {𝛼(𝑠)}π‘Žβˆˆ[0,𝑑]οΏ½οΏ½ and

(2) 𝑒(𝑑, π‘₯) = inf𝛼�𝐉̅ ��𝑋𝑠,𝐷π‘₯ (πœ”)οΏ½π‘Žβˆˆ[0,𝑑]; {𝛼(𝑠)}π‘Žβˆˆ[0,𝑑]οΏ½ + 𝑒 �𝑑, 𝑋𝑑,𝐷π‘₯ (πœ”)οΏ½οΏ½, where 𝛼(𝑑) ∈ π‘ˆ ⊊ 𝑅𝑛.

Here 𝐉̅ = πΈπ›ΊοΏ½βˆ« �𝑔�𝑋0𝑑 𝑠,𝐷π‘₯ (πœ”), 𝑠�� 𝑑𝑠� is the termination payoff: functional,𝛼(𝑑)is a control

and 𝑋𝑑,𝐷π‘₯ (πœ”) is some Markov process governed by some stochastic Ito’s equation driven

by a Brownian motion of the form

𝑋𝑑,𝐷π‘₯ (πœ”) = π‘₯ + ∫ 𝑓 �𝑋0𝑑 𝑠,𝐷π‘₯ (πœ”), 𝛼(𝑠)�𝑑𝑠 + βˆšπ·π‘Š(𝑑, πœ”). (3)

Hereπ‘Š(𝑑, πœ”) is the Brownian motion. Traditionally the function 𝑒(𝑑, π‘₯) has been computed by way of solving the associated Bellman equation, for which various numerical techniques mostly variations of the finite difference scheme have been developed. Another approach, which takes advantage of the recent developments in computing technology and allows one to construct the function𝑒(π‘₯, 𝑑) by way of backward induction governed by Bellman’s principle such that described in [1]. In paper [1] Equation (3) is approximated by an equation with affine coefficients which admits an explicit solution in terms of integrals of the exponential Brownian motion. Using Colombeau approach proposed in paper [2], [3],[4] we have replaced Equation (3) by Colombeau-Ito’s equation [4-6]: �𝑋𝑑,𝐷,πœ€π‘₯,πœ€ β€²(πœ”, πœ›)οΏ½ πœ€β€² = π‘₯ + οΏ½οΏ½ π‘“πœ€β€²οΏ½π‘‹π‘‘,𝐷,πœ€β€² π‘₯,πœ€ (πœ”, πœ›), 𝛼(𝑠)οΏ½ 𝑑 0 π‘‘π‘ οΏ½πœ€β€² +√𝐷 �∫ 𝑀0𝑑 πœ€β€²(𝑠, πœ›)𝑑𝑠� πœ€β€²+ βˆšπœ€οΏ½π‘Š(𝑑, πœ”)οΏ½πœ€β€². Here πœ€, πœ€β€²βˆˆ (0,1],πœ” ∈ 𝛺

1, πœ› ∈ 𝛺2, 𝛺1∩ 𝛺2 = βˆ…, where 𝑀(𝑑, πœ”) is the white noise on ℝ𝑛

i.e., 𝑀(𝑑, πœ”) = 𝑑 𝑑𝑑⁄ π‘Š(𝑑, πœ”) almost surelyin 𝐷′and 𝑀

(3)

noiseon ℝ𝑛i.e., 𝑀

πœ€β€²(𝑑, πœ›) = βŒ©π‘€(𝑑, πœ”), πœ™πœ€β€²(𝑠 βˆ’ 𝑑)βŒͺ, and πœ™πœ€β€² is a model delta net [2], [4].

Fortunately in contrast with Equation (3) one can solve Equation (4) without any approximation using strong large deviations principleof Non-Freidlin-Wentzelltype [5],[6],[7].

Statement of the novelty and uniqueness of the proposed idea: A new approach, which is proposed in this paper allows one to construct the Bellman function𝑣(𝑑, π‘₯) and optimal control 𝛼(𝑑, π‘₯) directly, i.e., without any reference to the Bellman equation, by way of using strong large deviations principle for the solutionsColombeau-Ito’s SDE (CISDE).

2. Proposed Approach

Letℭ𝑖 = (Ω𝑖, πšΊπ‘–, 𝐏𝑖), 𝑖 = 1,2 be a probability spaces such that: Ξ©1∩ Ξ©2 = βˆ…. Let us

consider m-persons Colombeau-Ito differential gameπΆπΌπ·πΊπ‘š;𝑇(𝒇, π’ˆ, π’š, 𝐺𝑛(ℝ𝑛), β„­1, β„­2),

withthe termination payoff functional for the i-th player is: οΏ½π‰Μ…πœ€πœ€β€²,𝑗� πœ€β€²=𝐄𝛺1𝐄𝛺2��∫ 𝑔0𝑇 πœ€β€²,𝑖�π‘₯𝑑,𝐷,πœ€π‘₯,πœ€ β€²(πœ”, πœ›), 𝜢(𝑑), 𝑑, πœ€οΏ½π‘‘π‘‘οΏ½ πœ€β€²οΏ½ + +𝐄𝛺1𝐄𝛺2οΏ½οΏ½βˆ‘ οΏ½π‘₯𝑇,𝐷,πœ€π‘₯,πœ€ β€²;𝑖(πœ”, πœ›) βˆ’ 𝑦𝑖� 2 𝑛 𝑖=1 οΏ½ πœ€β€²οΏ½ (1)

and with stochastic nonlinear dynamics: �𝒙̇𝑑,𝐷,πœ€π‘₯0,πœ€β€²(πœ”, πœ›)οΏ½ πœ€β€² = οΏ½π’‡πœ€β€²οΏ½π’™π‘‘,𝐷,πœ€β€² π‘₯,πœ€ (πœ”, πœ›), βˆšπ·π‘€ πœ€β€²(𝑑, πœ›), 𝜢(𝑑), 𝑑, πœ€οΏ½οΏ½ πœ€β€² + βˆšπœ€οΏ½π’˜(𝑑, πœ”)οΏ½πœ€β€² (2) πœ€, πœ€β€²βˆˆ (0,1],πœ” ∈ 𝛺 1, πœ› ∈ 𝛺2. Here βˆ€π‘‘ ∈ [0, 𝑇]: οΏ½π’™πœ€β€²(𝑑)οΏ½πœ€β€² ∈ ℝ�𝑛; 𝒙0,𝐷,πœ€π‘₯0,πœ€β€²(πœ”, πœ›) = 𝒙0 ∈ ℝ𝑛, βˆ€πœΊ ∈ (0,1]:𝒇 = [(π’‡πœ€β€²)πœ€β€²], π’ˆ = [(π’ˆπœ€β€²)πœ€β€²]; 𝒇(𝒙,∘,∘,∘,∘), π’ˆ(𝒙,∘,∘,∘) ∈ 𝐺𝑛(ℝ𝑛), 𝜢(𝑑) = {𝛼1(𝑑), … , π›Όπ‘š(𝑑)}; 𝛼𝑖(𝑑) ∈ π‘ˆπ‘– ⊊ β„π‘˜π‘–, 𝑖 = 1, … , π‘š,

And m–persons Colombeau-Ito differential game

πΆπΌπ·πΊπ‘š;𝑇(𝒇, π’ˆ, π’š, 𝐺𝑛(ℝ𝑛), 𝜷(𝑑), 𝝋(𝑑), β„­1, β„­2) with imperfect measurements and with

imperfect information about the system [5], [6]. The corresponding stochastic nonlinear dynamics is: �𝒙̇𝑑,𝐷,πœ€π‘₯0,πœ€β€²(πœ”, πœ›)οΏ½ πœ€β€² = οΏ½π’‡πœ€β€²οΏ½π’™π‘‘,𝐷,πœ€β€² π‘₯0,πœ€ (πœ”, πœ›), βˆšπ·π‘€ πœ€β€²(𝑑, πœ›), 𝝋(𝑑), 𝛼 �𝑑, π‘₯𝑑,𝐷,πœ€π‘₯0,πœ€β€² + 𝜷(𝑑)οΏ½ , 𝑑, πœ€οΏ½οΏ½ πœ€β€²+

(4)

+βˆšπœ€οΏ½π‘€(𝑑, πœ”)οΏ½πœ€β€²; πœ€, πœ€β€²βˆˆ (0,1], πœ” ∈ Ξ©1, πœ› ∈ Ξ©2 and the playoff for the i-th player is: οΏ½π‰Μ…πœ€πœ€β€²,𝑗� πœ€β€²=𝐄𝛺1𝐄𝛺2��∫ π‘”πœ€β€²,𝑖�𝒙𝑑,𝐷,πœ€β€² π‘₯0,πœ€ (πœ”, πœ›), 𝛼�𝑑, 𝜷(𝑑)οΏ½, 𝑑, πœ€οΏ½ 𝑇 0 π‘‘π‘‘οΏ½πœ€β€²οΏ½ + +𝐄𝛺1𝐄𝛺2οΏ½οΏ½βˆ‘ οΏ½π‘₯𝑇,𝐷,πœ€π‘₯0,πœ€β€²;𝑖(πœ”, πœ›) βˆ’ 𝑦𝑖� 2 𝑛 𝑖=1 οΏ½ πœ€β€²οΏ½. (3) Here𝜷(𝑑) =(𝛽1(𝑑), … , 𝛽𝑛(𝑑)), 𝝋(𝑑) = οΏ½πœ‘1(𝑑), … , πœ‘π‘›(𝑑)οΏ½and βˆ€π‘‘ ∈ [0, 𝑇]: οΏ½π‘₯πœ€β€²(𝑑)οΏ½πœ€β€² ∈ ℝ�𝑛; π‘₯ 0,𝐷,πœ€π‘₯0,πœ€β€²(πœ”, πœ›) = π‘₯0, βˆ€πœΊ ∈ (0,1]: 𝒇 = [(π’‡πœ€β€²)πœ€β€²], π’ˆ = [(π’ˆπœ€β€²)πœ€β€²]; 𝒇(π‘₯,∘,∘,∘,∘,∘), π’ˆ(π‘₯,∘,∘,∘) ∈ 𝐺𝑛(ℝ𝑛)or 𝒇(π‘₯,∘,∘,∘,∘), π’ˆ(π‘₯,∘,∘) ∈ 𝐺𝑃,π‘Ÿπ‘› (𝐸), 𝛼(𝑑) = {𝛼1(𝑑), … , π›Όπ‘š(𝑑)}; 𝛼𝑖(𝑑) ∈ π‘ˆ ⊊ β„π‘˜π‘–, 𝑖 = 1, … , π‘š, 𝜷(𝑑) = {𝛽1(𝑑), … , 𝛽𝑛(𝑑)}, 𝝋(𝑑) = {πœ‘1(𝑑), … , πœ‘π‘›(𝑑)}.

Here ℝ is a field of the real numbers, 𝐺(ℝ𝑛) is the algebra of Colombeau generalized

functions [8],[9],[12] , 𝐺𝑛(ℝ𝑛) = 𝐺(ℝ𝑛) Γ— … Γ— 𝐺(ℝ𝑛), 𝐺

𝑃,π‘Ÿ(𝐸) =β„±β„ͺ𝑃,π‘Ÿπ‘ƒ,π‘Ÿ(𝐸)(𝐸)is the Colombeau

type algebra[13],[14], E is an appropriate algebra of functions, which is a locally convex vector space over field β‚΅, 𝐺𝑃,π‘Ÿπ‘› (𝐸) = 𝐺𝑃,π‘Ÿ(𝐸) Γ— … Γ— 𝐺𝑃,π‘Ÿ(𝐸), ℝ οΏ½ is the ring of Colombeau

generalized numbers [11], ℝ�𝒏 = ℝ� Γ— … Γ— ℝ�,𝑑 β†’ 𝛼

𝑖(𝑑)is the control chosen by the i-th

player, within a set of admissible control values π‘ˆπ‘–.

Here𝑑 ⟼ οΏ½οΏ½π‘₯𝑑,𝐷,πœ€π‘₯0,πœ€β€²;1(πœ”, πœ›)οΏ½

πœ€β€², … , οΏ½π‘₯𝑑,𝐷,πœ€β€²;𝑛

π‘₯0,πœ€ (πœ”, πœ›)οΏ½

πœ€β€²οΏ½isthe

trajectoryof the Equation (2). Optimal control problem for thei-th player is:

οΏ½π‰ΜΏπœ€πœ€β€²,𝑖�

πœ€β€² = οΏ½min𝛼𝑖(𝑑)βˆˆπ‘ˆπ‘–οΏ½max𝛼𝑗(𝑑)βˆˆπ‘ˆπ‘—π‰Μ…πœ€πœ€β€²,𝑗≠𝑖��

πœ€β€² . (4)

We remind now some classical definitions. Let us consider now Ito’s SDE:

𝑑𝒙𝑑= 𝒃(𝒙𝒕, 𝑑)𝑑𝑑 + βˆ‘π‘˜π‘Ÿ=1πœŽπ‘Ÿ(𝒙𝒕, 𝑑)π‘‘π‘Šπ‘Ÿ(𝑑, πœ”), (5)

𝒙0 = 𝒙0(πœ”), 𝒙 ∈ ℝ𝑛.

Theorem1.[15]-[16]. Let the vectors𝒃(𝒙, 𝑑), 𝝈(𝒙, 𝑑)be continuous functions of(𝒙, 𝑑)such that for some constants𝐷and 𝐢the following conditionshold:

(5)

‖𝒃(𝒙, 𝑑) βˆ’ 𝒃(π’š, 𝑑)β€– + βˆ‘ |πœŽπ‘˜π‘Ÿ=1 π‘Ÿ(𝒙, 𝑑) βˆ’ πœŽπ‘Ÿ(π’š, 𝑑)| ≀ 𝐷‖𝒙 βˆ’ π’šβ€–, (6)

‖𝒃(𝒙, 𝑑)β€– + βˆ‘ |πœŽπ‘˜π‘Ÿ=1 π‘Ÿ(𝒙, 𝑑)| ≀ 𝐢(1 + ‖𝒙‖). (7)

Then: (1) For every random variable𝒙(πœ”)independent of theprocessesπ‘Šπ‘Ÿ(𝑑, πœ”), π‘Ÿ =

1,2, … , π‘˜there exists a solution 𝒙𝑑of the Ito’s SDE(5)which is an almost surely

continuous Markov process and

(2)Two solutions𝒙𝑑,1and𝒙𝒕,2(πœ”)is unique up to equivalence: 𝐏�𝒙𝑑,1(πœ”) = 𝒙𝒕,2(πœ”)οΏ½ = 1,for

all 𝑑 ∈ [0, ∞) = 𝐼∞.

Remark1.[15],[17].It well known, that the boundedness assumption on 𝒃(𝒙, 𝑑)and 𝝈(𝒙, 𝑑)can be weakened, but somekind of restriction on the 𝒃(𝒙, 𝑑)and 𝝈(𝒙, 𝑑)is

necessary in order to guarantee the existence of a global solution i.e., a solution defined for all𝑑 ∈ [0, ∞).If we remove this condition of boundedness, then a solution of Ito’s SDE (5) does exist locally but, in general, blows up (or explodes) in finite time.

Definition1.Letℝ̀𝑛 = ℝ𝑛⋃{𝜟}be the one-point compactification of ℝ𝑛and𝑾̀𝑛 =

οΏ½π’˜|οΏ½0, ∞) βˆ‹ 𝑑 ↦ π’˜(𝑑) ∈ ℝ̀𝑛is continuous and such that

ifπ’˜(𝒕) = 𝜟, thenπ’˜(𝑑′) = 𝜟 for all 𝑑′β‰₯ 𝑑}.Let β„ŒοΏ½β„Μ€π‘›οΏ½be the 𝜎-field generated by Borel

cylinder sets. Forπ’˜ ∈ 𝑾̀𝑛we set

𝑒(π’˜) = inf{𝑑|π’˜(𝑑) = π›₯} (8) and call the explosion time of the trajectory π’˜(𝑑), 𝑑 ∈ [0, ∞).

Definition2.[15].By a solution𝒙𝑑(πœ”)of the equation(5)we mean a

�𝑾̀𝑛, β„ŒοΏ½β„Μ€π‘›οΏ½οΏ½- valued random variable defined on a probability

spaceβ„­ = (Ξ©, 𝚺, 𝐏)with areference family(πšΊπ‘‘)𝑑β‰₯0such that:

(i) there exists an n-dimensional(πšΊπ‘‘)-Brownian motion

𝑾(𝑑, πœ”) = οΏ½π‘Š1(𝑑, πœ”), … , π‘Šπ‘›(𝑑, πœ”)οΏ½ with 𝑾(0, πœ”) = 0,

(6)

(iii) if 𝑒(πœ”) = 𝑒�𝒙𝒕(πœ”)οΏ½ is the explosion time of 𝒙𝒕(πœ”)then for almost all πœ”,

𝒙𝑑(πœ”) βˆ’ 𝒙0(πœ”) =

= ∫ 𝒃(π’™πŸŽπ’• 𝒔(πœ”), 𝑑)𝑑𝑠 + βˆ‘π‘˜π‘Ÿ=1∫ 𝜎0𝑑 π‘Ÿ(𝒙𝒔(πœ”), 𝑑)π‘‘π‘Šπ‘Ÿ(𝑠, πœ”),(9)

for all 𝑑 ∈ οΏ½0, 𝑒(πœ”)οΏ½.

Theorem2.[15],[16].(1)Given β„Μ€π‘›βˆ’continuous𝒃(𝒙, 𝑑)and 𝝈(𝒙, 𝑑)consider the equation

(5).Then for any probability πœ‡ on�𝑾̀𝑛, β„ŒοΏ½β„Μ€π‘›οΏ½οΏ½with compact support, there exists a

solution

of (5) such thatthe law of 𝒙0(πœ”) coincides with πœ‡.

(2) Suppose𝒃(𝒙, 𝑑)and𝝈(𝒙, 𝑑)are locally Lipschitz continuous, i.e., for every 𝑁 > 0 there exists a constant𝐷𝑁> 0 such that

‖𝒃𝑛(𝒙, 𝑑)βˆ’π’ƒπ‘›(π’š, 𝑑)β€– + βˆ‘ οΏ½πœŽπ‘˜π‘Ÿ=1 π‘Ÿ,𝑛(𝒙, 𝑑) βˆ’ πœŽπ‘Ÿ,𝑛(π’š, 𝑑)οΏ½ ≀ 𝐷𝑁‖𝒙 βˆ’ π’šβ€– (10)

for every𝒙, π’š ∈ 𝑩𝑁,𝑩𝑡 = {𝒛|‖𝒛‖ ≀ 𝑁}.Then for any probability πœ‡ on�𝑾̀𝑛, β„ŒοΏ½β„Μ€π‘›οΏ½οΏ½with

compact support, there exists a solution

of (5) such thatthe law of 𝒙0(πœ”) coincides with πœ‡.

Theorem3.[16].Let𝒙𝑑,𝑛(𝑑), 𝑛 = 1,2, … be the solutions

of the Ito’s SDE’s

𝑑𝒙𝑑,𝑛 = 𝒃𝑛�𝒙𝑑,𝑛, 𝑑�𝑑𝑑 + βˆ‘π‘˜π‘Ÿ=1πœŽπ‘Ÿ,𝑛�𝒙𝑑,𝑛, π‘‘οΏ½π‘‘π‘Šπ‘Ÿ(𝑑, πœ”), (11)

𝒙0,𝑛 = 𝒙(πœ”)π‘₯ ∈ ℝ𝑛.

Assume that: (i) let the vectors 𝒃𝑛(𝒙, 𝑑), πˆπ‘›(𝒙, 𝑑)becontinuous functions of (𝒙, 𝑑)such

that for some constants𝐷and 𝐢 the following conditions hold

‖𝒃𝑛(𝒙, 𝑑) βˆ’ 𝒃𝑛(π’š, 𝑑)β€– + βˆ‘ οΏ½πœŽπ‘˜π‘Ÿ=1 π‘Ÿ,𝑛(𝒙, 𝑑) βˆ’ πœŽπ‘Ÿ,𝑛(π’š, 𝑑)οΏ½ ≀ 𝐷‖𝒙 βˆ’ π’šβ€–, (12)

(7)

(ii)𝐄[𝒙2(πœ”)] < ∞, (14)

(iii)βˆ€π‘ > 0:

limnβ†’βˆžsupβ€–x‖≀N�‖𝒃𝑛(𝒙, 𝑑) βˆ’ 𝒃0(π’š, 𝑑)β€– + βˆ‘ οΏ½πœŽπ‘˜π‘Ÿ=1 π‘Ÿ,𝑛(𝒙, 𝑑) βˆ’ πœŽπ‘Ÿ,0(π’š, 𝑑)οΏ½οΏ½ = 0. (15)

Then

limnβ†’βˆžsup0≀t≀T𝐄�𝒙𝑑,𝑛(πœ”) βˆ’ 𝒙𝑑,0(πœ”)οΏ½2 = 0. (16)

Corollary 1.Let 𝒙𝑑,𝑛(𝑑), 𝑛 = 1,2, … be the solutionsof the Ito’s SDE’s

𝑑𝒙𝑑,𝑛 = 𝒃𝑛�𝒙𝑑,𝑛, 𝑑�𝑑𝑑 + βˆ‘π‘˜π‘Ÿ=1πœŽπ‘Ÿ,π‘›π‘‘π‘Šπ‘Ÿ(𝑑, πœ”)𝒙0,𝑛 = 𝒙(πœ”)π‘₯ ∈ ℝ𝑛.(17)

Assume that: (i) Let the vectors 𝒃𝑛(𝒙, 𝑑),becontinuous functions of (𝒙, 𝑑) and πœŽπ‘› =

π‘π‘œπ‘›π‘ π‘‘ such that for some constants 𝐷and 𝐢 the following conditions hold ‖𝒃𝑛(𝒙, 𝑑) βˆ’ 𝒃𝑛(π’š, 𝑑)β€– ≀ 𝐷‖𝒙 βˆ’ π’šβ€– (18) ‖𝒃𝑛(𝒙, 𝑑)β€– + βˆ‘ οΏ½πœŽπ‘˜π‘Ÿ=1 π‘Ÿ,𝑛� ≀ 𝐢(1 + ‖𝒙‖), (19) (ii) 𝐄[𝒙2(πœ”)] < ∞, (20) (iii)βˆ€π‘ > 0: limnβ†’βˆžsupβ€–x‖≀N�‖𝒃𝑛(𝒙, 𝑑) βˆ’ 𝒃0(π’š, 𝑑)β€– + βˆ‘ οΏ½πœŽπ‘˜π‘Ÿ=1 π‘Ÿ,𝑛�� = 0. (21) Then limnβ†’βˆžsup0≀t≀T𝐄�𝒙𝑑,𝑛(πœ”) βˆ’ 𝒙𝑑,0(πœ”)οΏ½2 = 0.(22)

Here 𝒙𝑑,0(πœ”) is the solution of the ODE:

(8)

Remark 2.Note that Theorem 3 in fact asserts that under conditions (12)-(15)any solution𝒙𝑑(πœ”) of the Ito’s SDE (5) is continuously depend on functions 𝒃(𝒙, 𝑑)and𝝈(𝒙, 𝑑).

Note that the assumptions of the Lipschitz continuously (12) and boundedness (13) on 𝒃(𝒙, 𝑑) and 𝝈(𝒙, 𝑑)in the Theorem 3 cannot be weakened.

Theorem. Assume that: (1) Let 𝒙𝑑,𝑛(𝑑), 𝑛 = 1,2, … be the solutions of the Ito’s SDE’s

𝒙𝑑,𝑛 = 𝒃𝑛�𝒙𝑑,𝑛, 𝑑�𝑑𝑑 + πˆπ‘›οΏ½π’™π‘‘,𝑛, 𝑑�𝑑𝑾(𝑑, πœ”),

𝒙0,𝑛 = 𝒙(πœ”), π‘₯ ∈ ℝ𝑛.

and let𝒙�𝑑,𝑛(𝑑), 𝑛 = 1,2, … be the solutions of the Ito’s SDE’s

𝒙�𝑑,𝑛 = 𝒃�𝑛�𝒙�𝑑,𝑛, 𝑑�𝑑𝑑 + πˆοΏ½π‘›οΏ½π’™οΏ½π‘‘,𝑛, 𝑑�𝑑𝑾(𝑑, πœ”), 𝒙�0,𝑛 = 𝒙(πœ”), π‘₯ ∈ ℝ𝑛. Here πˆπ‘›οΏ½π’™π‘‘,𝑛, 𝑑�𝑑𝑾(𝑑, πœ”) = οΏ½ πœŽπ‘Ÿ,𝑛�𝒙𝑑,𝑛, π‘‘οΏ½π‘‘π‘Šπ‘Ÿ(𝑑, πœ”) π‘˜ π‘Ÿ=1 , πˆοΏ½π‘›οΏ½π’™οΏ½π‘‘,𝑛, 𝑑�𝑑𝑾(𝑑, πœ”) = οΏ½ πœŽοΏ½π‘Ÿ,𝑛�𝒙𝑑,𝑛, π‘‘οΏ½π‘‘π‘Šπ‘Ÿ(𝑑, πœ”) π‘˜ π‘Ÿ=1 . (2) The inequalities ‖𝒃𝑛(𝒙, 𝑑)β€– + β€–πˆπ‘›(𝒙, 𝑑)β€– ≀ 𝐾𝑛(1 + ‖𝒙‖), ‖𝒃𝑛(𝒙, 𝑑) βˆ’ 𝒃𝑛(π’š, 𝑑)β€– + β€–πˆπ‘›(𝒙, 𝑑) βˆ’ πˆπ‘›(𝒙, 𝑑)β€– ≀ 𝐾𝑛‖𝒙 βˆ’ π’šβ€–, �𝒃�𝑛(𝒙, 𝑑)οΏ½ + β€–πˆοΏ½π‘›(𝒙, 𝑑)β€– ≀ 𝐾𝑛(1 + ‖𝒙‖),

(9)

�𝒃�𝑛(𝒙, 𝑑) βˆ’ 𝒃�𝑛(π’š, 𝑑)οΏ½ + β€–πˆοΏ½π‘›(𝒙, 𝑑) βˆ’ πˆοΏ½π‘›(𝒙, 𝑑)β€– ≀ 𝐾𝑛‖𝒙 βˆ’ π’šβ€–,

�𝒃𝑛(𝒙, 𝑑) βˆ’ 𝒃�𝑛(𝒙, 𝑑)οΏ½ ≀ 𝛿1,𝑛‖𝒙‖,

β€–πˆπ‘›(𝒙, 𝑑) βˆ’ πˆπ‘›(𝒙, 𝑑)β€– ≀ 𝛿2,𝑛‖𝒙‖

where 0 ≀ 𝑑 ≀ 𝑇, is satisfied. Then the inequality

sup0≀𝑑≀𝑇𝐄 ��𝒙𝑑,π‘›βˆ’ 𝒙�𝑑,𝑛�2οΏ½ ≀ 𝑒𝐿𝑛�𝛿1,𝑛2 + 𝛿2,𝑛2 �𝐸 οΏ½οΏ½ �𝒙�𝑑,𝑛�2𝑑𝑑 𝑇

0 οΏ½

is satisfied.

Proof. See Appendix.

Remark 3.[17].If conditions(6)-(7)are valid only in every cylinder π‘ˆπ‘…Γ— 𝐼∞, with 𝐢 =

𝐢(𝑅), 𝐷 = 𝐷(𝑅), one can construct a sequenceof functions 𝒃𝑛(𝒙, 𝑑)andπˆπ‘›(𝒙, 𝑑)such that

for‖𝒙‖ < 𝑛

𝒃𝑛(𝒙, 𝑑) = 𝒃(𝒙, 𝑑), πˆπ‘›(𝒙, 𝑑) = 𝝈(𝒙, 𝑑), (24)

and therefore for each𝒃𝑛(𝒙, 𝑑), πˆπ’(𝒙, 𝑑)satisfy conditions(6)-(7)everywhere in ℝ𝑛. By

Theorem 1,there exists a sequence of Markov processes 𝒙𝑑,𝑛(πœ”)corresponding to the

functions𝒃𝑛(𝒙, 𝑑)and πˆπ’(𝒙, 𝑑).

Assumption1.Suppose now that the distribution of𝒙0(πœ”)hascompact support

inℝ𝑛.Then as, well known, that the first exit random times𝝉

π‘š(πœ”) of the

processes 𝒙𝑑,,π‘š(πœ”)from the set ‖𝒙‖ < 𝑛 are identical for π‘š β‰₯ 𝑛[15] [18],[19]. Let this

common value be𝝉𝑛(πœ”). Itis also clear that the processes themselves coincide up to

time𝝉𝑛(πœ”), i.e.

(10)

Orin the equivalent form

𝐏�sup0β‰€π‘‘β‰€πœπ‘›(πœ”)�𝒙𝑑,𝑛(πœ”) βˆ’ 𝒙𝒕,π‘š(πœ”)οΏ½ > 0οΏ½ = 0, π‘š > 𝑛. (27)

Definition1. (i) Let π‰βˆž(πœ”)denote the (finite or infinite) limit of the monotone

increasing sequence πœπ‘›(πœ”)as 𝑛 β†’ ∞. We call the random variable π‰βˆž(πœ”)the first exit

time from every bounded domain, or briefly the explosion time. (ii)We now define a new stochastic process 𝒙𝑑(πœ”)by setting[17]:

𝒙𝑑(πœ”) = 𝒙𝒕,𝑛(πœ”)for 𝑑 < 𝝉𝑛(πœ”). (28)

It well known, that this is always a Markov process for 𝑑 < 𝝉𝑛(πœ”)[18],[19].

We also can to define a new stochastic process 𝒙𝑑(πœ”)by setting

𝒙𝑑(πœ”) = 𝐏 βˆ’ limnβ†’βˆžπ’™π’•,𝑛(πœ”) (29)

If finite or infinite limit in RHS of Eq.(29) exist. (iii) In general case we set

�𝒙𝑑,πœΊβ€²(πœ”)οΏ½

πœΊβ€² = �𝒙𝑑,𝑛(πœ”)�𝑛, 𝑛 = 1

πœΊβ€². (30)

We note that the Colombeau-Ito’s equation �𝒙𝑑,πœΊβ€²(πœ”)οΏ½ πœΊβ€² βˆ’ �𝒙0,πœ€(πœ”)οΏ½πœΊβ€² = �∫ π’ƒοΏ½π’™πŸŽπ’• 𝒔,πœΊβ€²(πœ”), π‘ οΏ½π‘‘π‘ οΏ½πœΊβ€² + βˆ‘ �∫ 𝜎0𝑑 π‘ŸοΏ½π’™π’”,πœΊβ€²(πœ”), π‘‘οΏ½π‘‘π‘Šπ‘Ÿ(𝑠, πœ”)οΏ½ πœΊβ€² π‘˜ π‘Ÿ=1 (31)

is satisfied for all 𝑑 ∈ οΏ½0, π‰βˆž(πœ”)οΏ½.

(11)

𝐏𝑠,𝒙{𝝉

∞(πœ”) = ∞} = 1. (32)

Assumption2. We assume now

that:(1)βˆ€πœ– ∈ (0,1]:οΏ½π’ƒπœΊβ€²(𝒙, 𝑑, πœ–)οΏ½πœΊβ€² β‰œ �𝑏1,πœΊβ€²(𝒙, 𝑑, πœ–), … , 𝑏𝑛,πœΊβ€²(𝒙, 𝑑, πœ–)οΏ½πœΊβ€² ∈ 𝐺𝑛(ℝ𝑛) οΏ½or 𝐺𝑃,π‘Ÿπ‘› (𝐸)οΏ½,πœ– =

(πœ–1, … , πœ–π‘›), πœ– ∈ (0,1]:𝑛for all 𝑑 ∈ [0, ∞)and

(2) βˆ€πœ– ∈ (0,1]𝑛there exist infinite Colombeau constants �𝐢

πœ€πœ€β€²οΏ½πœ€β€²andοΏ½π·πœ€πœ€β€²οΏ½πœ€β€²such thatβˆ€πœ– ∈

(0,1]:

(i)(β€–π’ƒπœΊβ€²(𝒙, 𝑑, πœ–)β€–πŸ)πœΊβ€² ≀ οΏ½οΏ½πΆπœ€πœ€β€²οΏ½πœΊβ€²οΏ½ (1 + β€–π’™β€–πŸ), πœ€β€²βˆˆ (0,1],(33)

(ii)οΏ½οΏ½π’ƒπ’Š,πœΊβ€²(𝒙, 𝑑, πœ–) βˆ’ 𝒃𝑖,πœΊβ€²(π’š, 𝑑, πœ–)οΏ½οΏ½

πœΊπ’Šβ€² ≀ οΏ½οΏ½π·πœ€πœ€β€²οΏ½πœΊπ’Šβ€²οΏ½ ‖𝒙 βˆ’ π’šβ€–(34)

for all 𝑑 ∈ [0, ∞)and for all x∈ ℝ𝒏and for ally∈ ℝ𝑛.

Definition 2. [4] 1.Let β„­ = (𝛺, 𝛴. 𝑃)be a probability space. Let ℇ𝑅 be the space of nets οΏ½π‘‹πœ€(πœ”)οΏ½πœ€of measurable functions on 𝛺.

Let ℇ𝑅𝑀 be the space of nets (π‘‹πœ€)πœ€ ∈ ℇ𝑅, πœ€ ∈ (0,1],with the

property that for almost all πœ” ∈ 𝛺 there exist constants π‘Ÿ, 𝐢 > 0 andπœ€0 ∈ (0,1] such that |(𝑋ℰ)β„°| ≀ πΆπœ€βˆ’π‘Ÿ, πœ€ ≀ πœ€0.

2.Let 𝑁𝑅 is the space of nets(𝑋ℰ)β„° ∈ ℇ𝑅, πœ€ ∈ (0,1],with the property that foralmost

all πœ” ∈ 𝛺and all 𝑏 ∈ ℝ+there exist constants𝐢 > 0and πœ€0 ∈ (0,1] such that |(𝑋ℰ)β„°| ≀

πΆπœ€π‘, πœ€ ≀ πœ€

0.Thedifferential algebra 𝐺𝑅of Colombeau generalized random variables is

thefactor algebra 𝐺𝑅 = ℇ𝑅/𝑁𝑅. Let us consider now a family�𝒙𝑑,πœ–,πœ€π‘₯0 ,πœ€β€²οΏ½

πœ€β€² of the solutions Colombeau-Ito’s SDE:

�𝑑𝒙𝑑,,πœ–,πœ€π‘₯0,πœ€β€²(πœ”)οΏ½ πœ€β€² = οΏ½π’ƒπœΊβ€²οΏ½π‘₯𝑑,,πœ–,πœ€β€² π‘₯0 ,πœ€(πœ”), 𝑑, πœ–οΏ½οΏ½ πœ€β€²+ βˆšπœ€οΏ½π‘‘π‘Ύ(𝑑, πœ”)οΏ½πœ€β€², (35) �𝒙0,πœ–,,πœ€π’™πŸŽ,𝜺 β€²οΏ½ πœ€β€² = οΏ½π’™πœΊβ€² 𝒙0(πœ”)οΏ½ πœΊβ€², ∈ 𝐺𝑅, �𝐄 �𝒙0,πœ–,πœ€β€² π‘₯0,πœ€ οΏ½οΏ½ πœ€β€² = 𝒙0 ∈ 𝑅 οΏ½ 𝑛, (36) 𝑑 ∈ [0, 𝑇], πœ€, πœ€β€²βˆˆ (0,1].

Here (i)𝑾(𝑑, πœ”) = οΏ½π‘Š1(𝑑, πœ”), … , π‘Šπ‘›(𝑑, πœ”)οΏ½ is n-dimensional Brownian motion,(ii)βˆ€π‘‘ ∈

(12)

polynomial on variable 𝒙 = (π‘₯1, … , π‘₯𝑛),i.e.

𝑏0,𝑖(𝒙, 𝑑, πœ–) = βˆ‘π›Ό,|𝛼|β‰€π‘Ÿπ‘0,𝑖𝛼 (𝑑, πœ–)π‘₯𝛼, (37)

𝛼 = (𝑖1, … , 𝑖𝑛), |𝛼| = βˆ‘π‘›π‘—=1𝑖𝑗 , 0 ≀ 𝑖𝑗 ≀ 𝑝, or

(iii)) βˆ€π‘‘ ∈ [0, 𝑇]:οΏ½π’ƒπœ€β€²(π‘₯, 𝑑, πœ€)οΏ½

πœ€β€² ∈ 𝐺𝑃,π‘Ÿπ‘› (𝐸), 𝒃0(𝒙, 𝑑, πœ–) ≑ π’ƒπœ€β€²=0(𝒙, 𝑑, πœ–): ℝ𝑛 β†’ ℝ𝑛isℝ-analytic

function on variable 𝒙 = (π‘₯1, … , π‘₯𝑛),i.e.

𝑏0,𝑖(𝒙, 𝑑, πœ–) = βˆ‘βˆžπ‘Ÿ=1βˆ‘π›Ό,|𝛼|β‰€π‘Ÿπ‘0,𝑖𝛼 (𝑑, πœ–)π‘₯𝛼, (38) 𝛼 = (𝑖1, … , 𝑖𝑛), |𝛼| = βˆ‘π‘›π‘—=1𝑖𝑗 , 0 ≀ 𝑖𝑗 ≀ 𝑝and (iv) limβ€–π’™β€–β†’βˆžβ€–π’ƒ0(𝒙, 𝑑, πœ–)β€– ‖𝒙‖ = βˆžβ„ , (v) 𝑏𝑖,πœ€β€²(𝒙(𝑑), 𝑑, πœ–) = 𝑏𝑖,0(π’™πœΊβ€²(𝑑), 𝑑, πœ–).(39) Here π’™πœΊβ€²(𝑑) = οΏ½π‘₯1,πœ€β€²(𝑑), … , π‘₯𝑛,πœ€β€²(𝑑)οΏ½and π‘₯𝑖,πœ€β€²(𝑑) = οΏ½ π‘₯𝑖,πœ€β€² (𝑑) = π‘₯𝑖(𝑑) 1+(πœΊβ€²)πŸπ’π’™π’ŠπŸπ’(𝑑), 𝑙 β‰₯ 1 or π‘₯𝑖,πœ€β€²(𝑑) = π‘₯𝑖(𝑑)πœƒπœ–π‘–[π‘₯𝑖(𝑑)]. (40) 𝑖 = 1, . . , 𝑛. Here πœƒπœ–π‘–[𝑧] ∈ 𝐢∞(ℝ), π‘ π‘’π‘π‘οΏ½πœƒπœ–π‘–[𝑧]οΏ½ βŠ† [βˆ’πœˆ(πœ–π‘–), 𝜈(πœ–π‘–)]

(13)

⎩ βŽͺ ⎨ βŽͺ ⎧ πœƒπœ–π‘–[𝑧] = 1 ↔ 𝑧 ∈ [βˆ’πœˆ1(πœ–π‘–), 𝜈1(πœ–π‘–)] ⊊ [βˆ’πœˆ(πœ–π‘–), 𝜈(πœ–π‘–)], πœƒπœ–π‘–[𝑧] = 0 ↔ 𝑧 ∈ ℝ\[βˆ’πœˆ(πœ–π‘–), 𝜈(πœ–π‘–)], 0 ≀ πœƒπœ–π‘–[𝑧] ≀ 1 ↔ 𝑧 ∈ [βˆ’πœˆ(πœ–π‘–), 𝜈(πœ–π‘–)]\[βˆ’πœˆ1(πœ–π‘–), 𝜈1(πœ–π‘–)].

Remark 5.By Theorem 1 for every Colombeau generalized random variableοΏ½π’™πœΊπ’™β€²0(πœ”)οΏ½

πœΊβ€², ∈ 𝐺𝑅 such that

�𝐄 �𝒙0,πœ€π‘₯0,πœ€β€²οΏ½οΏ½

πœ€β€² = 𝒙0 ∈ 𝑅 �𝑛,and independent of the processes π‘Š1(𝑑, πœ”), … , π‘Šπ‘›(𝑑, πœ”) there exist

Colombeau generalized stochastic process �𝒙𝑑,πœ–,πœ€π’™πŸŽ,πœΊβ€²(πœ”)οΏ½ πœΊβ€², πœ€ β€²βˆˆ (0,1], such that�𝒙 0,πœ–,πœΊβ€² π’™πŸŽ,𝜺 (πœ”)οΏ½ πœΊβ€² = οΏ½π’™πœΊβ€² 𝒙0(πœ”)οΏ½ πœΊβ€²,and �𝒙𝑑,,πœ–,πœ€π’™πŸŽ,πœΊβ€²(πœ”)οΏ½

πœΊβ€²is the solution of the Colombeau-Ito’s SDE (35)-(36),which is an almost

surely continuous Colombeau generalized stochastic process and is unique up to equivalence

�𝐏 ��𝒙𝒕,,𝝐,πœΊπ’™πŸŽ,πœΊβ€²,1(πœ”) βˆ’ 𝒙𝒕,,𝝐,πœΊπ’™πŸŽ,πœΊβ€²,2(πœ”)οΏ½ > 0οΏ½οΏ½

πœΊβ€² = 0, for all 𝑑 ∈ [0, ∞).

Remark 6.One can to construct a sequence of Colombeau generalized functions οΏ½π’ƒπœ€β€²,𝑛(π‘₯, 𝑑, πœ–)οΏ½πœ€β€² such that for β€–π‘₯β€– < 𝑛:

π’ƒπœ€β€²,𝑛(π‘₯, 𝑑, πœ–) = π’ƒπœΊβ€²(π‘₯, 𝑑, πœ–), πœ€β€²βˆˆ (0,1], πœ– ∈ (0,1]𝑛,

and therefore for each π’ƒπœ€β€²,𝑛(π‘₯, 𝑑, πœ–), satisfy conditions (18)-(19) everywhere in ℝ𝑛. By

Theorem 1, there exists a sequence of Colombeau generalized stochastic processes �𝒙𝑑,,πœ€π‘₯0,πœ€β€²,𝑛(πœ”)οΏ½

πœ€β€²corresponding to Colombeau generalized functionsοΏ½π’ƒπœ€β€²,𝑛(π‘₯, 𝑑, πœ–)οΏ½πœ€β€². Suppose

now that for each πœ€β€²βˆˆ (0,1], πœ– ∈ (0,1]𝑛 the distribution of 𝒙 0,πœ€β€²

π‘₯0 (πœ”) has compact support

in ℝ𝑛. Then there exit times of the processes 𝒙

𝑑,,πœ–,πœ€π‘₯0,πœ€β€²,π‘š(πœ”), πœ€β€², πœ€ ∈ (0,1], from the set

β€–π‘₯β€– < 𝑛 are identical for π‘š β‰₯ 𝑛. Let this common value beπ‰πœ€β€²,𝑛(πœ”, πœ–). It is also clear

that the processes�𝒙𝑑,,πœ–,πœ€π‘₯0,πœ€β€²,𝑛(πœ”)οΏ½

πœ€β€²and �𝒙𝑑,,πœ–,πœ€β€²,π‘š π‘₯0,πœ€ (πœ”)οΏ½

πœ€β€²themselves coincide up to

timeοΏ½π‰πœ€β€²,𝑛(πœ”, πœ–)οΏ½

(14)

�𝐏 οΏ½sup0β‰€π‘‘β‰€π‰πœ€β€²,𝑛(πœ”,πœ–)�𝒙π‘₯𝑑,,,πœ–,πœ€0,πœ€ β€²,π‘š(πœ”) βˆ’ 𝒙𝑑,,,πœ–,πœ€π‘₯0,πœ€ β€²,𝑛(πœ”)οΏ½ > 0 οΏ½οΏ½

πœΊβ€² = 0, (41)

for all π‘š β‰₯ 𝑛.

Definition3. (i) Let π‰πœΊβ€²(πœ”, πœ€, πœ–), πœ€β€², πœ– ∈ (0,1]𝑛denote the (finite or infinite) limit of the

monotone increasing sequence π‰πœ€β€²,𝑛(πœ”, πœ€, πœ–)as 𝑛 β†’ ∞. We call the generalized random

variable οΏ½π‰πœΊβ€²(πœ”, πœ€, πœ–)οΏ½πœ€β€², πœ€β€²βˆˆ (0,1]the first exit time of the sample function from every

bounded domain, or briefly thegeneralizedexplosion time.

(ii) We now define Colombeau generalized stochastic process �𝒙𝒕,,∈,πœΊπ’™πŸŽ,πœΊβ€²(πœ”)οΏ½

πœΊβ€²by setting

𝒙𝒕,,𝝐,πœΊπ’™πŸŽ,πœΊβ€²(πœ”) = 𝒙𝒕,,𝝐,πœΊπ’™πŸŽ,πœΊβ€²,𝑛(πœ”)for 𝑑 = 𝑑(πœ”) < π‰πœΊβ€²,𝑛(πœ”, πœ€, πœ–). (42)

(iii)That this is always a Markov process for 𝑑 = 𝑑(πœ”) < οΏ½π‰πœΊβ€²,𝑛(πœ”, πœ€, πœ–)οΏ½ πœΊβ€².

(iv) Colombeau generalized stochastic process �𝒙𝒕,,𝝐,πœΊπ’™πŸŽ,πœΊβ€²(πœ”)οΏ½

πœΊβ€²defined by

setting (42) on the random generalized interval οΏ½0, οΏ½π‰πœΊβ€²,𝑛(πœ”, πœ€, πœ–)οΏ½ πœΊβ€²οΏ½is regular ,if for any 𝑠 < ∞, 𝒙 ∈ ℝ𝑛, 𝝐 ∈ (0,1]𝑛:

(𝐏𝑠,𝒙{𝝉

πœΊβ€²(πœ”, πœ€, πœ–) = ∞})πœΊβ€² = 1, πœ€β€²βˆˆ (0,1](43)

(vi) Colombeau generalized stochastic process �𝒙𝒕,𝝐,,πœΊπ’™πŸŽ,πœΊβ€²(πœ”)οΏ½

πœΊβ€², defined by setting (42)

is a strongly regular if for any 𝑠 < ∞, 𝒙 ∈ ℝ𝑛,πœ€β€²βˆˆ [0,1], 𝝐 ∈ (0,1]𝑛:

(𝐏𝑠,,𝒙{𝝉

πœΊβ€²(πœ”, πœ€, πœ–) = ∞})πœΊβ€² = 1. (44)

Remark7.Wenote that: (iii) does not imply (iv).

Proposition1.Assume that Colombeau generalized stochastic process�𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²(πœ”)οΏ½

πœΊβ€²

defined by setting (42)is a strongly regular. Then (1)βˆ€π, 𝝐 ∈ (0,1]𝑛, βˆ€π›Ώ, 𝛿 > 0:

(15)

limπœΊβ€²β†’0𝐄 ��𝒙𝑑,,πœ–,πœ€π’™πŸŽ,πœΊβ€²(πœ”) βˆ’ 𝒙𝑑,πœ–,πœ€π’™πŸŽ,πœΊβ€²=0(πœ”)οΏ½ 𝟐 οΏ½ = 0. (45.a) limπœΊβ€²β†’0𝐏 ��𝒙𝑑,,πœ–,πœ€π’™πŸŽ,πœΊβ€²(πœ”) βˆ’ 𝒙𝑑,πœ–,πœ€π’™πŸŽ,πœΊβ€²=0(πœ”)οΏ½ > 𝛿� = 0. (45.b) (2)βˆ€π›Ώ, 𝛿 > 0: limπœΊβ€²β†’0,𝝐→0𝐄 ��𝒙𝑑,,πœ–,πœ€π’™πŸŽ,πœΊβ€²(πœ”) βˆ’ 𝒙𝑑,πœ–=0,πœ€π’™πŸŽ,𝜺 β€²=0(πœ”)οΏ½ 2 οΏ½ = 0. (45.c) limπœΊβ€²β†’0,𝝐→0𝐏 ��𝒙𝑑,,πœ–,πœ€π’™πŸŽ,πœΊβ€²(πœ”) βˆ’ 𝒙𝑑,πœ–=0,πœ€π’™πŸŽ,𝜺 β€²=0(πœ”)οΏ½ > 𝛿� = 0. (45.d)

Proof. Immediately follows fromTheoremA1.(I) (see appendix A)anddefinitions1,3. Let us consider now a family �𝒙𝑑,πœ–,πœ€π‘₯0 ,πœ€β€²(πœ”)οΏ½

πœ€β€² of the solutions of the Colombeau SDE:

�𝑑𝒙𝑑,πœ–,,πœ€π‘₯0,πœ€β€²(πœ”)οΏ½ πœ€β€²,πœ– = οΏ½π’ƒπœΊβ€²,𝝐�𝒙𝑑,,πœ–,πœ€β€² π‘₯0 ,πœ€ (πœ”), 𝑑, πœ”οΏ½οΏ½ πœ€β€²+ βˆšπœ€π‘‘π‘Ύ(𝑑, πœ”), (46) �𝒙0,πœ€π‘₯0,πœ€β€²οΏ½ πœ€β€² = 𝒙0 ∈ 𝑅 �𝑛, 𝑑 ∈ [0, 𝑇], πœ€, πœ€β€², ∈ (0,1],πœ– ∈ (0,1]. 𝑛

Here 𝑾(𝑑) is n-dimensional Brownian motion,

andβˆ€πœ– ∈ (0,1]𝑛, βˆ€π‘‘ ∈ [0, 𝑇] andfor almost al Ο‰ ∈ Ξ© :οΏ½π’ƒπœ€β€²,πœ–(π‘₯, 𝑑, πœ”)οΏ½πœ€β€² ∈ 𝐺𝑛(ℝ𝑛), 𝒃0,0(βˆ™, 𝑑) ≑

π’ƒπœ€β€²=0,πœ–=0(βˆ™, 𝑑, πœ”): ℝ𝑛 β†’ ℝ𝑛is a polynomialvector-function on a variable 𝒙 = (π‘₯1, … , π‘₯𝑛) i.e.,

𝑏𝑖,0,0(𝒙, 𝑑) = βˆ‘π›Ό,|𝛼|β‰€π‘Ÿπ‘π‘–,0,0𝛼 (𝑑)π‘₯𝛼, 𝛼 = (𝑖1, … , 𝑖𝑛), |𝛼| = βˆ‘π‘›π‘—=1𝑖𝑗 , 0 ≀ 𝑖𝑗 ≀ 𝑝, and

𝑏𝑖,πœ€β€²,πœ–(𝒙(𝑑), 𝑑, πœ”) = 𝑏𝑖,0,0οΏ½π’™πœ€β€²,πœ–(𝑑, πœ”), 𝑑�. (47)

(16)

π‘₯π’Š,πœ€β€²,πœ–(𝑑, πœ”) = π‘₯𝑖(𝑑) 1+πœ€β€²π‘₯ 𝑖2𝑙(𝑑)+πœ€β€²οΏ½πœ–π‘–βˆ« πœƒ0𝑑 πœ–π‘–[π‘₯𝑖(𝜏)]π‘₯𝑖2𝑙(𝜏)π‘‘πœ+βˆšπ›Ώπ‘Šπ‘–(𝑑)οΏ½ 2 , (48) 𝑖 = 1, . . , 𝑛. Now we let 𝑒𝑖(𝑑) = πœ–π‘–βˆ« πœƒπœ–π‘–[π‘₯𝑖(𝜏)]π‘₯𝑖2𝑙(𝜏)π‘‘πœ + βˆšπ›Ώπ‘Šπ‘–(𝑑) 𝑑 0 (49)

and rewrite Eq.(46) of the canonical Colombeau-Ito form: �𝑑𝒙𝑑,,πœ€π‘₯0,πœ€β€²,πœ–(πœ”)οΏ½ πœ€β€²,πœ– = οΏ½π’ƒπœ€β€²,πœ–οΏ½π’™π‘‘,,πœ€β€²,πœ– π‘₯0 ,πœ€ (πœ”), 𝒖 𝑑,,πœ€β€²,,πœ– 𝛿 (πœ”), 𝑑�� πœ€β€² + βˆšπœ€π‘‘π‘Ύ(𝑑, πœ”), 𝒖𝑑,,πœ€β€²,πœ€,πœ– 𝛿 (πœ”) = �𝑒1,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”), … , 𝑒𝑛,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”)οΏ½, (50) �𝑑𝑒𝑖,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”)οΏ½ πœ€β€² = πœ–π‘–οΏ½πœƒπœ–π‘–οΏ½π‘₯𝑖,𝑑,πœ€π‘₯0 ,𝛿′,πœ–(πœ”)οΏ½ οΏ½π‘₯𝑖,𝑑,,πœ€π‘₯0 ,𝛿′,πœ–(πœ”)οΏ½ 2𝑙 οΏ½ πœ€β€² + βˆšπ›Ώπ‘‘π‘Šπ‘–(𝑑), (51) 𝑖 = 1, … , 𝑛, οΏ½π‘₯0,πœ€π‘₯0,πœ€β€² οΏ½ πœ€β€² = 𝒙0 ∈ 𝑅 οΏ½ 𝑛, 𝑑 ∈ [0, 𝑇], πœ€, πœ€β€², πœ–, 𝛿 ∈ (0,1].

Theorem3.Let us consider a pair of the Colombeau-Ito’s SDE: �𝑑𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,πœ‡(πœ”)οΏ½ πœΊβ€² = οΏ½π’ˆπœΊβ€² πœ‡ �𝒙 𝑑,,πœΊβ€² π’™πŸŽ ,𝜺(πœ”), 𝑑�� πœΊβ€² + βˆšπœ€οΏ½π‘‘π‘Ύ(𝑑, πœ”)οΏ½πœΊβ€², (52) �𝒙0,πœΊπ’™πŸŽ,πœΊβ€²,πœ‡οΏ½ πœΊβ€² = 𝒙0 ∈ 𝑅 οΏ½ π‘š, 𝑑 ∈ [0, 𝑇], πœ€, πœ€β€²βˆˆ (0,1], πœ‡ = 1,2. (53)

Assume now that:(1) Conditions (33) and (34)is satisfied. (2)For a given 𝑁 > 0, βˆ€π’™ ∈ ℝ𝑛such that ‖𝒙‖ ≀ 𝑁: π’ˆ

πœΊβ€²

1 (𝒙, 𝑑) = π’ˆ πœΊβ€² 2(𝒙, 𝑑).

Let𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,πœ‡(πœ”), πœ‡ = 1,2 be a pair of the solutions of the Colombeau- Ito’s SDE (52)-(53) and

letβ„±πœ€π‘,𝑑′,πœ‡(πœ”), πœ‡ = 1,2be a setβ„±πœ€π‘β€²,πœ‡(πœ”) = �𝑑|sup0≀𝑠≀𝑑�𝒙𝒔,,πœΊπ’™πŸŽ,πœΊβ€²,πœ‡(πœ”)οΏ½ ≀ 𝑁�.We let nowπ‰πœ€π‘β€²,πœ‡(πœ”) =

sup�𝑑|𝑑 ∈ β„±πœ€,πœ€π‘β€²,πœ‡(πœ”)οΏ½.

(17)

(i) ποΏ½π‰πœ€,πœ€π‘ β€²,1(πœ”) = π‰πœ€,πœ€π‘ β€²,2(πœ”)οΏ½ = 1 and

(ii) 𝐏 οΏ½sup0≀𝑠≀𝝉1�𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,1(πœ”) βˆ’ 𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,2(πœ”)οΏ½ = 0οΏ½ = 1.

Proof. A proof of this statement, complete similarly, to a classical case. For example see[15],chapt.2, subsect.6,theorem2.

Let us rewrite now Eq.(50)-Eq.(51) in the next form (with πœƒπœ–π‘– ≑ 1)

�𝒙𝑑,,πœ€π‘₯0,πœ€β€²,πœ–(πœ”, 𝛿)οΏ½ πœ€β€²,πœ– = 𝒙0 + �∫ 𝒃0𝑑 πœ€β€²,πœ–οΏ½π’™π‘₯0 ,πœ€πœ,,πœ€β€²,πœ–(πœ”, 𝛿), π’–πœ,,πœ€π›Ώ β€²,πœ–(πœ”), 𝜏� π‘‘πœοΏ½ πœ€β€²+ βˆšπœ€π‘Ύ(𝑑, πœ”), (54) 𝒖𝑑,,πœ€π›Ώ β€²,πœ–(πœ”) = �𝑒1,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”), … , 𝑒𝑛,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”)οΏ½, �𝑒𝑖,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”)οΏ½ πœ€β€² = πœ–π‘–οΏ½βˆ« οΏ½π‘₯𝑖,𝜏,,πœ€β€²,,πœ– π‘₯0 ,πœ€ (πœ”, 𝛿)οΏ½2π‘™π‘‘πœ 𝑑 0 οΏ½πœ€β€² + βˆšπ›Ώπ‘Šπ‘–(𝑑), (55) 𝑖 = 1, … , 𝑛,

Let 𝐺𝑁(π’š), π’š ∈ ℝ𝑛bea function: (i) 𝐺𝑁(π’š) = π’šifβ€–π’šβ€– ≀ 𝑁(ii)𝐺𝑁(π’š) = 0ifβ€–π’šβ€– > 𝑁.

We set now π’ƒπœΊπ‘β€²,𝝐(𝒙, 𝒖, 𝑑) = π’ƒπœΊβ€²,𝝐(𝐺𝑁(𝒙), 𝐺𝑁(𝒖), 𝑑). LetοΏ½π’šπ‘‘,,πœ€π‘₯0,πœ€β€²,πœ–(πœ”, 𝛿, 𝑁)οΏ½ πœ€β€² = ��𝒙𝑑,,πœ€β€²,πœ– π‘₯0,πœ€ (πœ”, 𝛿, 𝑁)οΏ½ πœ€β€², 𝒖𝑑,,πœ€β€²,,πœ– 𝛿 (πœ”, 𝑁)οΏ½

be a family of the solution of the Colombeau-Ito’s SDE: �𝒙𝑑,,πœ€π‘₯0,πœ€β€²,πœ–(πœ”, 𝛿, 𝑁)οΏ½ πœ€β€² = 𝒙0+ �∫ π’ƒπœ€β€²,πœ– 𝑁 �𝒙 𝜏,,πœ€π‘₯0 ,πœ€β€²,πœ–(πœ”, 𝛿, 𝑁), π’–πœ,,πœ€π›Ώ β€²,πœ–(πœ”), 𝜏, 𝑁� 𝑑 𝑑 0 πœοΏ½πœ€β€²+ (56) +βˆšπœ€π‘Ύ(𝑑, πœ”), 𝒖𝑑,,πœ€π›Ώ β€²,πœ–(πœ”, 𝑁) = �𝑒1,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”, 𝑁), … , 𝑒𝑛,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”, 𝑁)οΏ½, �𝑒𝑖,𝑑,,πœ€π›Ώ β€²,πœ–(πœ”, 𝑁)οΏ½ πœ€β€² = πœ–π‘–οΏ½βˆ« �𝐺𝑁�π‘₯𝑖,𝜏,,πœ€π‘₯0 ,πœ€β€²,,πœ–(πœ”, 𝛿, 𝑁)οΏ½οΏ½ 2𝑙 π‘‘πœ 𝑑 0 οΏ½ πœ€β€² + +βˆšπ›Ώπ‘Šπ‘–(𝑑), (57)

(18)

𝑖 = 1, … , 𝑛,

Definition4.(1) Let οΏ½οΏ½π’šπ‘‘,,πœ€π‘₯0,πœ€β€²,πœ–(πœ”, 𝛿, 𝑁)οΏ½

πœ€β€²,πœ–οΏ½π‘=1 ∞

be a sequence of the solution of the Colombeau-Ito’s SDE(54)- (55). Letβ„±πœ€,πœ€π‘β€²,πœ–(πœ”, 𝛿) be a set

β„±πœ€,πœ€π‘β€²,πœ–(πœ”, 𝛿) = �𝑑|sup0β‰€π‘ β‰€π‘‘οΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,πœ–(πœ”, 𝛿, 𝑁)οΏ½ ≀ 𝑁�. (58)

(2) We let now

π‰πœ€,πœ€π‘ β€²,πœ–(πœ”, 𝛿) = sup�𝑑|𝑑 ∈ β„±πœ€,πœ€π‘β€²,πœ–(πœ”)οΏ½, (59)

π‰πœ€,πœ€βˆžβ€²,πœ–(πœ”, 𝛿) = limπ‘β†’βˆžπ‰πœ€,πœ€π‘ β€²,πœ–(πœ”, 𝛿).(60)

(3) Letπ’šοΏ½π’•,,πœΊπ’™πŸŽ,πœΊβ€²,πœ–(πœ”, 𝛿)be a net of the stochastic processes defined

by setting

π’šοΏ½π’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿) = π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿, 𝑁)iff 𝑑 < π‰πœΊ,πœΊπ‘β€²,𝝐(πœ”, 𝛿)(61)

(4) LetοΏ½οΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½

πœΊβ€²οΏ½be Colombeaugeneralized stochastic process defined by setting

οΏ½οΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½

πœΊβ€²οΏ½ = οΏ½οΏ½π’šοΏ½π‘‘,,πœ€β€²,πœ–

π’™πŸŽ,𝜺 (πœ”, 𝛿)οΏ½

πœΊβ€²οΏ½. (62)

Remark5.We note that according to the Theorem3 βˆ€π‘€(𝑀 β‰₯ 𝑁)one obtain 𝐏 οΏ½sup0≀𝑑≀𝝉

𝜺,,πœΊβ€²,𝝐

𝑡 (πœ”)οΏ½π’šπ’™π’•,,𝜺𝟎,πœΊβ€²,𝝐(πœ”, 𝛿, 𝑁) βˆ’ π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿, 𝑀)οΏ½ > 0οΏ½ = 0,

Therefore definitions (61)-(62) is correct. Definition5.LetοΏ½π’šπ‘‘,πœ€π‘₯0 ,πœ€β€²,πœ–οΏ½

πœΊβ€²be a family of the solutions Colombeau-Ito’s SDE (56)-(57).

(1) A familyοΏ½π’šπ‘‘,πœ€π‘₯0 ,πœ€β€²,πœ–οΏ½

πœΊβ€², πœ€, πœ€β€²βˆˆ (0,1],πœ– ∈ (0,1]

(19)

οΏ½limπ’„β†’βˆžπ οΏ½οΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½ > 𝑐��

πœΊβ€² = 0.(63)

Or in the next equivalent form οΏ½ποΏ½π‰πœΊ,πœΊβˆžβ€²,𝝐(πœ”, 𝛿) = ∞�� πœΊβ€² = 1. (64) (2) A familyοΏ½π’šπ’•,πœΊπ’™πŸŽβ€² ,𝜺,𝝐� πœΊβ€²is a strongly regular if βˆ€πœ€β€², πœ€β€² ∈ [0,1], βˆ€πœ–, πœ– ∈ (0,1] 𝑛: οΏ½limπ’„β†’βˆžπ ��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½ > 𝑐�� πœΊβ€² = 0.(65)

or in the next equivalent form: βˆ€πœ€β€², πœ€β€²βˆˆ [0,1], βˆ€πœ–, πœ– ∈ (0,1] 𝑛:

οΏ½ποΏ½π‰πœΊ,πœΊβˆžβ€²,𝝐(πœ”, 𝛿) = ∞��

πœΊβ€² = 1.(66)

Definition6. LetοΏ½π’šπ‘‘,πœ€π‘₯0 ,πœ€β€²,πœ–οΏ½

πœΊβ€²bea family of the solutions Colombeau-Ito’s SDE (56)-(57). A

family οΏ½π’šπ‘‘,πœ€π‘₯0 ,πœ€β€²,πœ–οΏ½ πœΊβ€², πœ€, πœ€ β€²βˆˆ οΏ½0,1], πœ– ∈ (0,1] 𝑛isanon-regular if βˆƒπ‘‘β€²βˆ€π‘‘ β‰₯ 𝑑′: οΏ½lim π’„β†’βˆžπ ��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½ > 𝑐�� πœΊβ€² β‰  0. (67)

Or in the next equivalent form οΏ½ποΏ½π‰πœΊ,πœΊβˆžβ€²,𝝐(πœ”, 𝛿) < ∞��

πœΊβ€² = 1. (68)

Proposition2. Assume that Colombeau generalized stochastic process οΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½

(20)

defined by setting (62) is a strongly regular. Then (1) limπœΊβ€²β†’0 πœ–β†’0,lim𝛿→0𝐄 οΏ½οΏ½π’šπ’•,,πœΊβ€²,𝝐 π’™πŸŽ,𝜺 (πœ”, 𝛿) βˆ’ π’š 𝒕 π’™πŸŽ,𝜺(πœ”)�𝟐� = 0 (69.a) (2) βˆ€πœŽ > 0: limπœΊβ€²β†’0 πœ–β†’0,lim𝛿→0𝐏 οΏ½οΏ½π’šπ‘‘,,πœΊβ€²,𝝐 π’™πŸŽ,𝜺 (πœ”, 𝛿) βˆ’ π’š 𝑑 π’™πŸŽ,𝜺(πœ”)οΏ½ > 𝜎� = 0. (69.b) Here:π’šπ‘‘π’™πŸŽ,𝜺(πœ”) = π’š 𝑑,,πœΊβ€²=0,πœ–=0 π’™πŸŽ,𝜺 (πœ”, 𝛿 = 0).

Proof. Immediately follows fromTheorem3 andTheoremA.1 (see appendix A). Proposition3.LetοΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½

πœΊβ€² = ��𝒙𝑑,πœ€β€²,πœ– π‘₯0 ,πœ€οΏ½

πœ€β€², �𝒖𝑑,,πœ€β€²,πœ–

𝛿 (πœ”)οΏ½

πœ€β€²οΏ½be a family of the solutions

Colombeau-Ito’s SDE (56)-(57)withπœƒπœ–[𝑧] ≑ 1.A family οΏ½π’šπ‘‘,πœ€π‘₯0 ,πœ€β€²,πœ–οΏ½

πœΊβ€², πœ€, πœ€β€²βˆˆ (0,1],𝝐 ∈ (0,1] 𝑛is

regular.

Proof. Assume that: process οΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½

πœΊβ€²is a non-regular. Therefore(𝐏 𝑠,,𝒙{𝝉 πœΊβ€²(πœ”, πœ€) < ∞})πœΊβ€² = 1 and consequently �𝐏𝑠,,π’™οΏ½π’š π‰πœΊβ€²(πœ”,πœ€),,πœ€β€²,πœ– π’™πŸŽ,𝜺 (πœ”, 𝛿) = ∞�� πœΊβ€² > 0. (70)

But the other hand from Eq.(56)-Eq. (57) we obtain

�𝒙𝝉 πœΊβ€²(πœ”,πœ€),,πœ€β€²,πœ– π‘₯0,πœ€ (πœ”, 𝛿)οΏ½ πœ€β€² = 𝒙0 + �∫0π‰πœΊβ€²(πœ”,πœ€)π’ƒπœ€β€²,πœ–οΏ½π’™π’—,,πœ€π‘₯0 ,πœ€β€²,πœ–(πœ”, 𝛿), 𝒖𝑣,,πœ€π›Ώ β€²,πœ–(πœ”), 𝑣, πœ€οΏ½ 𝑑𝑣� πœ€β€² + + οΏ½βˆšπœ€π‘Ύ(π‰πœΊβ€²(πœ”, πœ€), πœ”)οΏ½ πœ€β€², (71) 𝒖𝑣,,πœ€π›Ώ β€²,πœ–(πœ”) = �𝑒1,𝑣,,πœ€π›Ώ β€²,πœ–(πœ”), … , 𝑒𝑛,𝑣,,πœ€π›Ώ β€²,πœ–(πœ”)οΏ½, �𝑒𝑖,𝝉 πœΊβ€²(πœ”,πœ€),,πœ€β€²,πœ– 𝛿 (πœ”)οΏ½ πœ€β€² = πœ– �∫ οΏ½π‘₯𝑖,𝑣,,πœ€β€²,,πœ– π‘₯0 ,πœ€ (πœ”, 𝛿)οΏ½2𝑙𝑑𝑣 π‰πœΊβ€²(πœ”,πœ€) 0 οΏ½πœ€β€²+ οΏ½βˆšπ›Ώπ‘Šπ‘–οΏ½π‰πœΊβ€²(πœ”, πœ€)οΏ½οΏ½πœ€β€², 𝑖 = 1, … , 𝑛, (72)

(21)

From (70) and Eq. (71)-Eq. (72) we obtain �𝐏0,,π‘₯0οΏ½οΏ½ 𝒃 πœ€β€²,πœ–οΏ½π’™π‘£,,πœ€π‘₯0 ,πœ€β€²,πœ–(πœ”, 𝛿), 𝒖𝑣,,πœ€π›Ώ β€²,πœ–(πœ”), 𝑣, πœ€οΏ½ 𝑑 πœπœ€β€²(πœ”,πœ€) 0 𝑣 = βˆžοΏ½οΏ½πœΊβ€² = 0, and therefore �𝐏0,,π‘₯0�𝒙 π‰πœΊβ€²(πœ”,πœ€),,πœ€β€²,πœ– π‘₯0,πœ€ (πœ”, 𝛿) = ∞�� πœ€β€² = �𝐏0,,π‘₯0�𝒙0+ οΏ½ 𝒃 πœ€β€²,πœ–οΏ½π’™π‘£,,πœ€π‘₯0 ,πœ€β€²,πœ–(πœ”, 𝛿), 𝒖𝑣,,πœ€π›Ώ β€²,πœ–(πœ”), 𝑣, πœ€οΏ½ 𝑑 πœπœ€β€²(πœ”,πœ€) 0 𝑣 + βˆšπœ€π‘Ύ(π‰πœΊ β€²(πœ”, πœ€), πœ”) = ∞�� πœΊβ€² = 0, �𝐏0,,0�𝑒 𝑖,π‰πœΊβ€²(πœ”,πœ€),,πœ€β€²,πœ– 𝛿 (πœ”) = ∞�� πœ€β€² = �𝐏𝑠,,π’™οΏ½πœ– οΏ½ οΏ½π‘₯ π’Š,𝒗,,πœΊπ’™πŸŽ ,πœΊβ€²,,πœ–(πœ”, 𝛿)οΏ½ πŸπ’ 𝑑𝑣 + βˆšπ›Ώπ‘Šπ‘–οΏ½π‰πœΊβ€²(πœ”, πœ€)οΏ½ π‰πœΊβ€²(𝝎,𝜺) 0 = βˆžοΏ½οΏ½πœΊβ€² = 0. Thus �𝐏𝑠,,π’™οΏ½π’š π‰πœΊβ€²(πœ”,πœ€),,πœ€β€²,πœ– π’™πŸŽ,𝜺 (πœ”, 𝛿) = ∞�� πœΊβ€² = 0.

But this is the contradiction. This contradiction completed the proof.

Definition7.CISDE(35)-(36) is ℝ οΏ½-dissipative if there exist

Lyapunov candidate functionοΏ½π‘‰πœ€β€²(𝒙, 𝑑)οΏ½πœ€β€²: ℝ �𝑛× [0, 𝑇] β†’ ℝ οΏ½and positive infinite

Colombeau constants 𝐢� = [(πΆπœ€β€²)πœ€β€²] ∈ ℝ οΏ½+,

π‘ŸΜƒ = [(π‘Ÿπœ€β€²)πœ€β€²] ∈ ℝ οΏ½+,such that:

(1)βˆ€πœ€β€²βˆˆ (0,1] βˆΆπ‘‰

βˆ—,πœ€β€² = limπ‘…β†’βˆžοΏ½inf ‖𝐱‖>π‘…π‘‰πœ€β€²(𝒙, 𝑑)οΏ½ = ∞, and

(22)

οΏ½οΏ½π‘‰Μ‡πœ€β€²(π’™πœ€β€², 𝑑; π’ƒπœ€β€²)οΏ½ πœ€β€²οΏ½ ≀ 𝐢̃ οΏ½οΏ½π‘‰πœ€β€²(π’™πœ€β€², 𝑑)οΏ½πœ€β€²οΏ½ (73) is satisfied.Here οΏ½οΏ½π‘‰Μ‡πœ€β€²(π’™πœ€β€², 𝑑; π’ƒπœ€β€²)οΏ½ πœ€β€²οΏ½ ≑ οΏ½οΏ½πœ•π‘‰πœ€β€²(π’™πœ•π‘‘πœ€β€², 𝑑)οΏ½ πœ€β€²οΏ½ + οΏ½ οΏ½οΏ½ πœ•π‘‰πœ€β€²(π’™πœ€β€², 𝑑) πœ•π‘₯𝑖,πœ€β€² 𝑏𝑖,πœ€β€²(π’™πœ€β€², 𝑑)οΏ½ πœ€β€² οΏ½ 𝑛 𝑖=1 . (74)

Or in the next equivalent form:

CISDE (35)-(36) is ℝ οΏ½-dissipative if there existLyapunovcandidate

functionοΏ½π‘‰πœ€β€²(𝒙, 𝑑)οΏ½πœ€β€²: ℝ �𝑛× [0, 𝑇] β†’ ℝ οΏ½and positive infinite Colombeau constants 𝐢� =

[(πΆπœ€β€²)πœ€β€²] ∈ ℝ οΏ½+,

π‘ŸΜƒ = [(π‘Ÿπœ€β€²)πœ€β€²] ∈ ℝ οΏ½+, such that:

(1)βˆ€πœ€β€²βˆˆ (0,1] βˆΆπ‘‰

βˆ—,πœ€β€² = limπ‘…β†’βˆžοΏ½inf ‖𝐱‖>π‘…π‘‰πœ€β€²(𝒙, 𝑑)οΏ½ = ∞, and

(2β€²)βˆ€πœ€β€²βˆˆ (0,1]βˆ€π’™ πœ€β€²[(π’™πœ€β€² ∈ ℝ𝑛) ∧ (β€–π’™πœ€β€²β€– β‰₯ π‘Ÿπœ€β€²)]the inequality οΏ½π‘‰Μ‡πœ€β€²(π’™πœ€β€², 𝑑; π’ƒπœ€β€²)οΏ½ πœ€β€² ≀ ((πΆπœ€β€²)πœ€β€²)οΏ½π‘‰πœ€β€²(π’™πœ€β€², 𝑑)οΏ½πœ€β€² (75) is satisfied. Here οΏ½π‘‰Μ‡πœ€β€²(π‘₯πœ€β€², 𝑑; π‘πœ€β€²)οΏ½πœ€β€² ≑ οΏ½πœ•π‘‰πœ€β€²πœ•π‘‘οΏ½π’™πœ€β€²,𝑑�� πœ€β€² + οΏ½βˆ‘ πœ•π‘‰πœ€β€²οΏ½π’™πœ€β€²,𝑑� πœ•π‘₯πœ€β€² 𝑏𝑖,πœ€β€²(π’™πœ€β€², 𝑑) 𝑛 𝑖=1 οΏ½ πœ€β€². (76)

Definition 8. CISDE (35)-(36) isa stronglyℝ οΏ½-dissipative if

Lyapunov candidate functionοΏ½π‘‰πœ€β€²(𝒙, 𝑑)οΏ½

πœ€β€²: ℝ �𝑛× [0, 𝑇] β†’ ℝ οΏ½,

πœ€β€² ∈ [0,1]andpositive finite Colombeau constants

𝐢� = [(πΆπœ€β€²)πœ€β€²] ∈ ℝ οΏ½+, π‘ŸΜƒ = [(π‘Ÿπœ€β€²)πœ€β€²] ∈ ℝ οΏ½+,such that:

(1)βˆ€πœ€β€²βˆˆ (0,1] βˆΆπ‘‰

βˆ—,πœ€β€² = limπ‘Ÿβ†’βˆžοΏ½inf β€–xβ€–>π‘Ÿπ‘‰πœ€β€²(𝒙, 𝑑)οΏ½ = ∞,and(2)βˆ€[(π’™πœ€β€²)πœ€β€²]([(β€–π’™πœ€β€²β€–)πœ€β€²] β‰₯ π‘ŸΜƒ) the

inequality

οΏ½οΏ½π‘‰Μ‡πœ€β€²(π’™πœ€β€², 𝑑; π’ƒπœ€β€²)οΏ½

(23)

is satisfied. Here οΏ½οΏ½π‘‰Μ‡πœ€β€²(π’™πœ€β€², 𝑑; π’ƒπœ€β€²)οΏ½πœ€β€²οΏ½ ≑ οΏ½οΏ½πœ•π‘‰πœ€β€²(π’™πœ•π‘‘πœ€β€², 𝑑)οΏ½ πœ€β€² οΏ½ + οΏ½ οΏ½οΏ½πœ•π‘‰πœ€πœ•π‘₯β€²(π’™πœ€β€², 𝑑) 𝑖,πœ€β€² 𝑏𝑖,πœ€β€²(π’™πœ€β€², 𝑑)οΏ½ πœ€β€² οΏ½ 𝑛 𝑖=1 . (78) Proposition4.Let�𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”)οΏ½

πœΊβ€²) be generalized stochastic process satisfying

Colombeau-Ito’s SDE(35)-(36) on the time interval[𝑠, 𝑇]and οΏ½π‰πœΊβ€²,π‘ˆ(πœ”, πœ–)οΏ½

πœΊβ€²- is a

generalizedrandomvariableequaltothetimeatwhichthesamplefunctionofthegeneralized process �𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”)οΏ½

πœΊβ€²) first leaves the bounded neighborhood π‘ˆ, and

letοΏ½π‰πœΊβ€²,π‘ˆ(πœ”, 𝑑, πœ–)οΏ½πœΊβ€² = οΏ½minοΏ½π‰πœΊβ€²,π‘ˆ(πœ”, πœ–), π‘‘οΏ½οΏ½πœΊβ€².Supposemoreover

thatβˆ€πœ€β€²βˆˆ (0,1] ∢𝐏 �𝒙 𝒔,,πœΊβ€²,𝝐 π’™πŸŽ,𝜺 (πœ”) ∈ π‘ˆοΏ½ = 1.Then �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π‰π’™πœΊβ€²,π‘ˆπŸŽ,𝜺(πœ”,𝑑,πœ–),,πœΊβ€²,𝝐(πœ”), π‰πœΊβ€²,π‘ˆ(πœ”, 𝑑, πœ–)οΏ½ βˆ’ π‘‰πœ€β€²οΏ½π’™π’”,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”), 𝑠��� πœ€β€² = = �𝐄 οΏ½οΏ½ π‘‰Μ‡πœ€β€²οΏ½π’™π‘’,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”), 𝑒� 𝑑𝑒 π‰πœΊβ€²,π‘ˆ(πœ”,𝑑,πœ–) 𝑠 οΏ½οΏ½πœ€β€².

Proof. Similarly as the proof of the corresponding classical result, see [17]Lemma 3.2. Theorem4. (1) Assume that: (i)for CISDE (35)-(36) the inequalities(33)and(34) is satisfied and (ii)CISDE (35)-(36) isℝ οΏ½-dissipative.

Then (1) Colombeau generalized stochastic process �𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”)οΏ½ πœΊβ€², πœ€

β€²βˆˆ (0,1] , πœ– ∈

(0,1] 𝑛defined by setting (42) is regular, and (2) the inequality �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”), 𝑑���

πœ€β€² ≀ �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π’•πŸŽ,,πœΊβ€²,𝝐 π’™πŸŽ,𝜺 (πœ”), 𝑑

0οΏ½οΏ½οΏ½πœ€β€²exp[(πΆπœ€β€²)πœ€β€²(𝑑 βˆ’ 𝑑0)](79)

is satisfied.

Proof.(1) From (76) it follows that the Colombeau generalized functionοΏ½π‘Šπœ€β€²(π’™πœ€β€², 𝑑)οΏ½ πœ€β€² =

οΏ½π‘‰πœ€β€²(π’™πœ€β€², 𝑑)οΏ½

(24)

Proposition4,forοΏ½π‰πœΊβ€²,𝑛(πœ”, 𝑑, πœ–)οΏ½ πœΊβ€² = οΏ½minοΏ½π‰πœΊβ€²,𝑛(πœ”, πœ–), π‘‘οΏ½οΏ½πœΊβ€²wehave �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π‰π’™πœΊβ€²,,𝝐,π‘›πŸŽ,𝜺 (πœ”,𝑑,πœ–),,πœΊβ€²(πœ”), π‰πœΊβ€²,𝑛(πœ”, 𝑑, πœ–)οΏ½ expοΏ½βˆ’(πΆπœ€β€²)πœ€β€²οΏ½π‰πœΊβ€²,𝑛(πœ”, 𝑑, πœ–) βˆ’ 𝑑0οΏ½οΏ½οΏ½οΏ½ πœ€β€² βˆ’ �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π’• 𝟎,,πœΊβ€²,𝝐 π’™πŸŽ,𝜺 (πœ”), 𝑑 0οΏ½οΏ½οΏ½πœ€β€² = �𝐄 οΏ½βˆ«π’•π‰πœΊβ€²,𝑛(πœ”,𝑑,πœ–)π‘ŠΜ‡πœ€β€²οΏ½π’™π‘’,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”), 𝑒� 𝑑𝑒 𝟎 οΏ½οΏ½πœ€β€² ≀ 0. (80)

This, together with the inequalitiesοΏ½π‰πœΊβ€²,𝑛(πœ”, 𝑑, πœ–)οΏ½

πœΊβ€² ≀ 𝑑, οΏ½π‘‰πœ€β€²(π’™πœ€β€², 𝑑)οΏ½ πœ€β€² β‰₯ 0, implies �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π‰ πœΊβ€²,πœ–,𝑛(πœ”,𝑑),,πœΊβ€² π’™πŸŽ,𝜺 (πœ”), 𝝉 πœΊβ€²,π‘ˆ(πœ”, 𝑑, πœ–)οΏ½οΏ½οΏ½ πœ€β€² ≀ �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™οΏ½π’• 𝟎,πœ–,πœΊβ€² π’™πŸŽ,𝜺 (πœ”), 𝑑 0οΏ½οΏ½οΏ½πœ€β€²exp[(πΆπœ€β€²)πœ€β€²(𝑑 βˆ’ 𝑑0)] (81)

From (81) one derive the estimate

οΏ½ποΏ½π‰πœΊβ€²,𝑛(πœ”, πœ–) < π‘‘οΏ½οΏ½πœ€β€² ≀ ≀ exp[(πΆπœ€β€²)πœ€β€²(𝑑 βˆ’ 𝑑0)] �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™οΏ½π’•πŸŽ,,πœΊβ€² π’™πŸŽ,𝜺 (πœ”), 𝑑 0οΏ½οΏ½οΏ½πœ€β€² οΏ½inf‖𝐱‖β‰₯n,u>𝑑0π‘‰πœ€β€²(𝒙, 𝑒)οΏ½ πœΊβ€²

Letting 𝑛 β†’ ∞ and making use of theDefinition7we now get(64). (2)Assume that CISDE (35)-(36) is a strongly

ℝ

οΏ½-dissipative. Then(1) Colombeau generalized stochastic process = ��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²(πœ”)οΏ½ πœΊβ€²οΏ½ πœ€

β€²βˆˆ

[0,1], πœ– ∈ [0,1],𝑛defined by setting (42) is a strongly regular and (2) the inequality

�𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π’•,,πœΊπ’™πŸŽ,πœΊβ€²(πœ”), 𝑑��� πœ€β€² ≀ ≀ �𝐄 οΏ½π‘‰πœ€β€²οΏ½π’™π’• 𝟎,,πœΊβ€² π’™πŸŽ,𝜺 (πœ”), 𝑑 0οΏ½οΏ½οΏ½πœ€β€²exp[(πΆπœ€β€²)πœ€β€²(𝑑 βˆ’ 𝑑0)](82) is satisfied.

(25)

Theorem5.Weset nowπœƒππ’Š[𝒛] ≑ 1, 𝑖 = 1, … , 𝑛. For any solution

�𝒙𝑑,,πœ€π‘₯0,πœ€β€²,𝝐(πœ”, 𝛿)οΏ½

πœ€β€² = οΏ½π‘₯1,𝑑,,πœ€β€²,𝝐 π‘₯0,πœ€ , … , π‘₯

𝑛,𝑑,,πœ€π‘₯0,πœ€β€²,ποΏ½πœ€β€²

ofa stronglyℝ�–dissipative CISDE(46)-(48) and anyℝ-valued parametersπœ†1, … , πœ†π‘›, there

exist finite Colombeau constant 𝐢�′= ��𝐢

πœ€β€²β€²οΏ½πœ€β€²οΏ½ > 0, such that βˆ€π€[𝝀 = (πœ†1, … , πœ†π‘›)], the

inequality lim πœ€β†’0 πœ€β€²β†’0 πœ–β†’0 οΏ½πœ€β€²πœ€οΏ½β†’0 lim𝛿→0𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿) βˆ’ 𝝀� 𝟐 οΏ½ ≀ 𝐢̃′‖𝑼(𝑑, 𝝀)β€–πŸ (83)

is satisfied. Or in the next equivalent form: for a sufficiently smallπœ– β‰ˆ 0and for a sufficiently small πœ€ β‰ˆ 0, πœ€β€²β‰ˆ 0such that

πœ€β€² πœ€ β‰ˆ 0,the inequality οΏ½οΏ½lim𝛿→0𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿) βˆ’ 𝝀� 𝟐 οΏ½οΏ½ πœ€β€²οΏ½ ≀ 𝐢̃ ′‖𝑼(𝑑, 𝝀)β€–2 (84) is satisfied.

Here thevector-function 𝑼(𝑑, 𝝀) = (π‘ˆ1(t,𝝀),…,π‘ˆπ‘›(t,𝝀))is the solutionof the differential

master equation:

𝑼̇(𝑑, 𝝀)= 𝐉[π’ƒπŸŽ(𝝀, 𝑑)]𝑼(𝑑, 𝝀) + π’ƒπŸŽ(𝝀, 𝑑), 𝑼(0, 𝝀) = π’™πŸŽβˆ’ 𝝀, (85)

Here𝐉 = 𝐉[π’ƒπŸŽ(𝝀, 𝑑)] is a Jacobian i.e.,Jis 𝑛 Γ— 𝑛-matrix:

𝐉[π’ƒπŸŽ(𝝀, 𝑑)]= JοΏ½πœ•π’ƒπŸŽ,π’Š(𝒙, 𝑑)/πœ•π‘₯𝒋�𝒙=𝝀. (86)

Proof. We let now

𝒙𝑑,,πœ€π‘₯0,πœ€β€²,πβˆ’ 𝝀 = π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐. (86)

(26)

οΏ½π‘‘π’šπ‘‘,,πœ€π‘₯0,πœ€β€²,πœ–(πœ”, 𝛿)οΏ½ πœ€β€² = οΏ½π’ƒπœ€β€²,πœ–οΏ½π’šπ‘‘,,πœ€β€²,πœ– π‘₯0 ,πœ€ (πœ”, 𝛿) + 𝝀, 𝒖 𝑑,,πœ€β€²,πœ€,πœ– 𝛿 (πœ”), 𝑑, πœ€οΏ½οΏ½ πœ€β€² + +βˆšπœ€π‘‘π‘Ύ(𝑑, πœ”), 𝒖𝑑,,πœ€β€²,πœ€,πœ– 𝛿 (πœ”) = �𝑒1,𝑑,,πœ€π›Ώ β€²,πœ€,πœ–(πœ”), … , 𝑒𝑛,𝑑,,πœ€π›Ώ β€²,πœ€,πœ–(πœ”)οΏ½, (87) �𝑑𝑒𝑖,𝑑,,πœ€π›Ώ β€²,πœ€,πœ–(πœ”)οΏ½ πœ€β€² = πœ– οΏ½οΏ½π‘₯𝑖,𝑑,,πœ€β€²,,πœ€,πœ– π‘₯0 ,πœ€ (πœ”, 𝛿)οΏ½2𝑙� πœ€β€² + βˆšπ›Ώπ‘‘π‘Šπ‘–(𝑑), 𝑖 = 1, … , 𝑛, οΏ½π‘₯0,πœ€π‘₯0,πœ€β€²,πœ–οΏ½ πœ€β€² = 𝒙0 ∈ 𝑅 �𝑛, 𝑑 ∈ [0, 𝑇], πœ€, πœ€β€², πœ–, 𝛿 ∈ (0,1].

Thus we need to estimate the quantity lim πœ€β†’0 πœ€β€²β†’0 πœ–β†’0 οΏ½πœ€β€²πœ€οΏ½β†’0 lim𝛿→0π„π›€οΏ½οΏ½π’šπ’•,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½ 𝟐 οΏ½.

Application of the Theorem B.4 (see Appendix B) to Eq.(87) gives the inequality (83) directly.

Theorem 6.(Strong large deviations principle) [5],[7]. Assume that CISDE (35)-(36) is a stronglyℝ οΏ½-dissipative. Then:

(1) For any solution

�𝒙𝑑,,πœ€π‘₯0,πœ€β€²,𝝐(πœ”)οΏ½

πœ€β€² = οΏ½π‘₯1,𝑑,,πœ€β€²,𝝐 π‘₯0,πœ€ , … , π‘₯

𝑛,𝑑,,πœ€π‘₯0,πœ€β€²,𝝐� πœ€β€²

of a stronglyℝ�–dissipative CISDE(35)-(40) and anyℝ-valued parametersπœ† 1, … , πœ†π‘›, there

exist finite Colombeau constant 𝐢�′= ��𝐢

πœ€β€²β€²οΏ½πœ€β€²οΏ½ > 0, such that βˆ€π€[𝝀 = (πœ†1, … , πœ†π‘›)]the

inequality lim πœ€β†’0 πœ€β€²β†’0 πœ–β†’0 οΏ½πœ€β€²πœ€οΏ½β†’0 𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ 𝝀� 𝟐 οΏ½ ≀ 𝐢̃′‖𝑼(𝑑, 𝝀)β€–πŸ(8 8)

is satisfied. Or in the next equivalent form: for a sufficiently small πœ– β‰ˆ 0 and for a sufficiently small πœ€ β‰ˆ 0, πœ€β€²β‰ˆ 0 such that

πœ€β€²/πœ€ β‰ˆ 0,the inequality ��𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ 𝝀� 𝟐 οΏ½οΏ½ πœ€β€²οΏ½ ≀ 𝐢̃ ′‖𝑼(𝑑, 𝝀)β€–2.

(27)

is satisfied.

(2) For any solution

�𝒙𝑑,,πœ€π‘₯0,πœ€β€²,𝝐(πœ”)οΏ½

πœ€β€² = οΏ½π‘₯1,𝑑,,πœ€β€²,𝝐 π‘₯0,πœ€ , … , π‘₯

𝑛,𝑑,,πœ€π‘₯0,πœ€β€²,ποΏ½πœ€β€²

of a stronglyℝ�–dissipative CISDE(35)-(40) andanyℝ-valued parametersπœ† 1, … , πœ†π‘›, there

exist finite Colombeauconstant 𝐢�′= ��𝐢

πœ€β€²β€²οΏ½πœ€β€²οΏ½ > 0, such that βˆ€π€[𝝀 = (πœ†1, … , πœ†π‘›)]the

inequality

limπœ€β†’0𝐄𝛀��𝒙𝑑,,πœΊπ’™πŸŽ,πœΊβ€²=0,𝝐=0(πœ”) βˆ’ 𝝀� 𝟐

οΏ½ ≀ 𝐢̃′‖𝑼(𝑑, 𝝀)β€–πŸ (89)

issatisfied.Here the vector-function 𝑼(𝑑, 𝝀) = (π‘ˆ1(t,𝝀),…,π‘ˆπ‘›(t,𝝀))is the solutionof the

differential master equation:

𝑼̇(𝑑, 𝝀)= 𝐉[π’ƒπŸŽ(𝝀, 𝑑)]𝑼(𝑑, 𝝀) + π’ƒπŸŽ(𝝀, 𝑑), 𝑼(0, 𝝀) = π’™πŸŽβˆ’ 𝝀, (90)

where𝐉 = 𝐉[π’ƒπŸŽ(𝝀, 𝑑)]is aJacobian i.e.,Jis 𝑛 Γ— 𝑛-matrix:

𝐉[π’ƒπŸŽ(𝝀, 𝑑)]= JοΏ½πœ•π’ƒπŸŽ,π’Š(𝒙, 𝑑)/πœ•π‘₯𝒋�𝒙=𝝀. Proof1.Fromtheequality 𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ 𝝀� 𝟐 οΏ½ = 𝐄𝛀���𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ π’™π’™πŸŽ,πœΊπ’•,,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½ + + �𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿) βˆ’ 𝝀�� 𝟐 οΏ½, by using the triangle inequality, one obtain

�𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ 𝝀� 𝟐 οΏ½ ≀ �𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ 𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿)οΏ½ 𝟐 οΏ½ + +�𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”, 𝛿) βˆ’ 𝝀� 𝟐

οΏ½.Therefore statement (1) immediately follows fromTheoremA1 (see appendix A), Proposition 2andTheorem5.

(28)

𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²=0,πœ–=0(πœ”) βˆ’ 𝝀� 𝟐

οΏ½ = 𝐄𝛀���𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²=0,πœ–=0(πœ”) βˆ’ 𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”)οΏ½ + �𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ 𝝀�� 𝟐

οΏ½, byusingthetriangle inequality, one obtain

�𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²=0,πœ–=0(πœ”) βˆ’ 𝝀� 𝟐 οΏ½ ≀ �𝐄𝛀��𝒙𝒙𝒕,,𝜺𝟎,πœΊβ€²=0,πœ–=0(πœ”) βˆ’ 𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”)οΏ½ 𝟐 οΏ½ +�𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²,𝝐(πœ”) βˆ’ 𝝀� 𝟐

οΏ½.Therefore statement (2) immediately follows fromTheoremA1 (see appendix A), Proposition 1andstatement (1).

Remark.5. We note that in general case the inequality οΏ½οΏ½π›Ώπœ€β€²(𝑑)οΏ½

πœ€β€²οΏ½ ≑ οΏ½οΏ½limΞ΅β†’0𝐄Ω��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²(πœ”) βˆ’ 𝒙𝒕,,πœΊπ’™πŸŽ,𝜺=0β€² οΏ½ 𝟐

οΏ½οΏ½

πœ€β€²οΏ½ β‰  0

is satisfied, see Example 1. Example 1.Figures 1-2. π‘₯̇𝑑π‘₯0 ,πœ€ = βˆ’π‘Ž βˆ™ οΏ½π‘₯

𝑑π‘₯0 ,πœ€οΏ½3βˆ’ 𝑏 βˆ™ οΏ½π‘₯𝑑π‘₯0 ,πœ€οΏ½2βˆ’ 𝑐 βˆ™ π‘₯𝑑π‘₯0 ,πœ€βˆ’ 𝜎 βˆ™ 𝑑𝑛 βˆ’ (91)

βˆ’πœ’ βˆ™ π‘‘π‘šβˆ™ sin(Ξ© βˆ™ π‘‘π‘˜) + βˆšπœ€π‘€(𝑑) , π‘₯

0π‘₯0 ,πœ€ = π‘₯0.

From Eq.(91) and general differential master equation (90) one obtain the next linear differential master equation:

𝑒̇(𝑑) = βˆ’(3π‘Žπœ†2+ 2π‘πœ† + 𝑐)𝑒(𝑑) βˆ’ (π‘Ž βˆ™ πœ†3+ 𝑏 βˆ™ πœ†2+ 𝑐 βˆ™ πœ†) βˆ’ (92)

βˆ’πœŽ βˆ™ π‘‘π‘›βˆ’ πœ’ βˆ™ π‘‘π‘šβˆ™ sin(Ξ© βˆ™ π‘‘π‘˜) , 𝑒(0) = π‘₯ 0.βˆ’ πœ†.

From the differential master equation (92) one obtain the transcendental master equation:

Figure 1.The solution of the Equation (8) in a comparison with a corresponding solution 𝒙(𝒕) of the ODE (10).

(29)

Figure 2. 𝛿(r)versusR.

οΏ½π‘₯0 βˆ’ πœ†(𝑑)οΏ½expοΏ½βˆ’οΏ½3π‘Ž βˆ™ πœ†2(𝑑) + 2𝑏 βˆ™ πœ†(𝑑)οΏ½ βˆ™ 𝑑� βˆ’

βˆ’ ∫ [𝜎 βˆ™ πœπ‘‘ 𝑛+ πœ’ βˆ™ πœπ‘šβˆ™ sin(Ξ© βˆ™ πœπ‘˜) + π‘Ž βˆ™ πœ†3(𝑑) + 𝑏 βˆ™ πœ†2(𝑑)]

0 Γ— (93)

Γ— expοΏ½βˆ’οΏ½3π‘Ž βˆ™ πœ†2(𝑑) + 2𝑏 βˆ™ πœ†(𝑑)οΏ½ βˆ™ (𝑑 βˆ’ 𝜏)οΏ½π‘‘πœ = 0.

Example 1.Numerical simulation: Figures 1 and 2.

π‘Ž = 1, 𝑏 = 5, 𝑐 = 1, 𝜎 = πœ’ = βˆ’2, π‘š = 𝑛 = π‘˜ = 2, 𝛺 = 5, π‘₯0. = 0, 𝑇 = 5, 𝑅 = 𝑇/0.001. 𝛿(π‘Ÿ) = limπœ€β†’0𝐄 οΏ½οΏ½π‘₯𝑑π‘₯0 ,πœ€βˆ’π‘₯𝑑π‘₯0 ,πœ€=0οΏ½ 2 οΏ½. (94) π‘₯̇𝑑0 = βˆ’π‘Ž(π‘₯𝑑0)3βˆ’ 𝑏(π‘₯𝑑0)2βˆ’ 𝑐π‘₯𝑑0 βˆ’ 𝜎 βˆ™ π‘‘π‘›βˆ’ πœ’ βˆ™ π‘‘π‘šβˆ™ sin(Ξ© βˆ™ π‘‘π‘˜).(95)

Let β„­ = (Ξ©, 𝚺, 𝐏)be a probability space. Let us consider now

m–persons Colombeau-Ito’s stochastic differential gameCIDGπ‘š;𝑇(𝒇, π’ˆ, π’š, 𝐺𝑛(ℝ𝑛), β„­)with

nonlinear dynamics: �𝒙̇𝑑,πœ€π‘₯0,πœ€β€²(πœ”)οΏ½ πœ€β€² = οΏ½π’‡πœ€β€²οΏ½π’™π‘‘,πœ€β€² π‘₯0,πœ€(πœ”), 𝜢(𝑑), 𝑑�� πœ€β€²+ βˆšπœ€οΏ½π’˜(𝑑, πœ”)οΏ½πœ€β€²(96) Hereπœ€, πœ€β€²βˆˆ (0,1], πœ€ β‰ͺ 1; βˆ€π‘‘ ∈ [0, 𝑇]: �𝒙 πœ€β€²(𝑑)οΏ½ πœ€β€² ∈ 𝑅�𝑛; 𝒙0,πœ€π‘₯0,πœ€β€²(πœ”) = 𝒙0 ∈ ℝ𝑛, π’‡πœ€β€² = οΏ½π‘“πœΊβ€²,𝟏, … , π’‡πœΊβ€²,𝑛�, 𝒇 = [(π’‡πœ€β€²)πœ€β€²], π’ˆ = [(π’ˆπœ€β€²)πœ€β€²]; 𝒇(π‘₯,∘,∘), π’ˆ(π‘₯,∘,∘) ∈ 𝐺𝑛(ℝ𝑛), 𝜢(𝑑) = {𝛼 1(𝑑), … , π›Όπ‘š(𝑑)}; 𝛼𝑖(𝑑) ∈ π‘ˆπ‘– ⊊ β„π‘˜π‘–, 𝑖 = 1, … , π‘š.

Here𝑑 ⟼ 𝛼𝑖(𝑑), is the control chosen by the i-th player, within a set of admissible

(30)

i-th player is οΏ½π‰Μ…πœ€πœ€β€²,𝑖� πœ€β€²=𝐄 ��∫ 𝑔0𝑇 πœ€2β€²,𝑖�π‘₯𝑑,πœ€π‘₯0,πœ€β€² (πœ”), 𝜢(𝑑), 𝑑�𝑑𝑑� πœ€β€²οΏ½ + (97) +𝐄 οΏ½οΏ½οΏ½ οΏ½π‘₯𝑇,πœ€π‘₯,πœ€β€²;𝑖(πœ”) βˆ’ 𝑦𝑖� 2 𝑛 𝑖=1 οΏ½πœ€β€²οΏ½.

Definition 9.CIDGπ‘š;𝑇(𝒇, π’ˆ, π’š, 𝐺𝑛(ℝ𝑛), β„­) (96)-(97) is a strongly ℝ�–dissipative if

CISDE(96) is a stronglyℝ�–dissipative. Theorem.7. Suppose that:

(1)CIDGπ‘š;𝑇(𝒇, π’ˆ, π’š, 𝐺𝑛(ℝ𝑛), β„­)(96)-(97) isa stronglyℝ �–dissipative,

(2)𝒇0(∘, 𝜢, 𝑑) ≑ π’‡πœΊβ€²=0(∘, 𝜢, 𝑑): ℝ𝒏 β†’ ℝ𝒏is a polynomial

on a variable𝒙 = (π‘₯1, … , π‘₯𝑛)and a linear function on a

variable𝜢(𝑑) = {𝛼1(𝑑), … , π›Όπ‘š(𝑑)}i.e.,π‘“πŸŽ,π’Š(𝒙, 𝜢, 𝑑) =

βˆ‘π,|𝝁|β‰€π’“π‘“πŸŽ,π’Šπ (𝑑)𝒙𝝁+ βˆ‘π’=πŸπ’Ž 𝑐𝑙.𝑖(𝑑)𝛼𝑙(𝑑), 𝝁 = (π’ŠπŸ, … , π’Šπ’), |𝝁| = βˆ‘ 𝑖𝒏𝒋=𝟏 𝒋 , 0 ≀ π’Šπ’‹ ≀ 𝑝,

(3)π’ˆ0(∘, 𝜢, 𝑑) ≑ π’ˆπœΊβ€²=0(∘, 𝜢, 𝑑): ℝ𝒏 β†’ ℝ𝒏is a polynomial

on a variable𝒙 = (π‘₯1, … , π‘₯𝑛)and a linear function on a

variable𝜢(𝑑) = {𝛼1(𝑑), … , π›Όπ‘š(𝑑)}i.e.,π‘”πŸŽ,π’Š(𝒙, 𝜢, 𝑑) =

βˆ‘π,|𝝁|β‰€π’“π‘”πŸŽ,π’Šπ (𝑑)𝒙𝝁+ βˆ‘π’=πŸπ’Ž 𝑑𝑙,𝑖(𝑑)𝛼𝑙(𝑑), 𝝁 = (π’ŠπŸ, … , π’Šπ’), |𝝁| = βˆ‘ 𝑖𝒏𝒋=𝟏 𝒋 , 0 ≀ π’Šπ’‹ ≀ 𝑝.

Then For any solution

�𝒙𝑑,,πœ€π‘₯0,πœ€β€²; 𝜢�(𝑑)οΏ½ = οΏ½οΏ½π‘₯1,𝑑,,πœ€π‘₯0,πœ€β€², … , π‘₯𝑛,𝑑,,πœ€π‘₯0,πœ€β€²οΏ½ ; {𝛼1(𝑑), … , π›Όπ‘š(𝑑)}οΏ½(98)

of theCIDGπ‘š;𝑇(𝒇, π’ˆ, π’š, 𝐺𝑛(ℝ𝑛), β„­) (96) -(97) and any ℝ-valued

parameter𝑠 �𝝀(1), 𝝀(2)οΏ½ = οΏ½οΏ½πœ† 1 (1), … , πœ† 𝑛 (1)οΏ½, οΏ½πœ† 1 (2), … , πœ† π‘š

(2)οΏ½οΏ½there exist finite Colombeau

constant 𝐢̃′= ��𝐢

πœ€β€²β€²οΏ½πœ€β€²οΏ½ > 0, such thatβˆ€π€οΏ½π€ = �𝝀(𝟏), 𝝀(𝟐)οΏ½οΏ½ the inequalities

(1)lim πœ€β†’0 πœ€β€²β†’0 οΏ½πœ€β€²πœ€οΏ½β†’0 𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²(πœ”) βˆ’ 𝝀(𝟏)οΏ½ 𝟐 οΏ½ ≀ 𝐢̃′�𝑼�𝑑, 𝝀(𝟏)��𝟐, (2) lim πœ€β†’0 πœ€β€²β†’0 οΏ½πœ€β€²πœ€οΏ½β†’0 𝐄𝛀��𝒙𝒙𝑇,,𝜺𝟎,πœΊβ€²(𝝎) βˆ’ 𝝀(𝟐)οΏ½ 𝟐 οΏ½ ≀ 𝐢̃′�𝑽�𝑇, 𝝀(𝟏)��𝟐,(99.a)

(31)

(3) lim πœ€β†’0 πœ€β€²β†’0 οΏ½πœ€β€²πœ€οΏ½β†’0 𝐄𝛀��𝒙𝒙𝑇,,𝜺𝟎,πœΊβ€²(πœ”) βˆ’ π’šοΏ½ 𝟐 οΏ½ ≀ 𝐢̃′‖𝑽(𝑇, π’š)β€–πŸ, (4)lim πœ€β†’0 πœ€β€²β†’0 οΏ½πœ€β€²πœ€οΏ½β†’0 οΏ½π‰Μ…πœ€πœ€β€²,𝑗� ≀ |𝑉𝑖(𝑇, π’š, 0)| + ‖𝑼(𝑇, π’š)β€–2,𝑖 = 1, … , π‘š, (5)limπœ€β†’0𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²=0(πœ”) βˆ’ 𝝀(𝟏)οΏ½ 𝟐 οΏ½ ≀ 𝐢̃′�𝑼�𝑑, 𝝀(𝟏)��𝟐, (6)π₯π’π¦πœΊβ†’πŸŽπ„π›€οΏ½οΏ½π’™π’™π‘‡,,𝜺𝟎,πœΊβ€²=0(πœ”) βˆ’ 𝝀(𝟐)οΏ½ 𝟐 οΏ½ ≀ π‘ͺ�′�𝑽�𝑇, 𝝀(𝟏)��𝟐, (7) limπœ€β†’0𝐄𝛀��𝒙𝒙𝑇,,𝜺𝟎,πœΊβ€²=0(πœ”) βˆ’ π’šοΏ½ 𝟐 οΏ½ ≀ 𝐢̃′‖𝑽(𝑇, π’š)β€–πŸ, (8)limπœ€β†’0οΏ½π‰Μ…πœ€πœ€β€²=0,𝑗� ≀ |𝑉𝑖(𝑇, π’š, 0)| + ‖𝑼(𝑇, π’š)β€–2,𝑖 = 1, … , π‘š

issatisfied.Or in the next equivalent form: for a sufficiently small πœ€ β‰ˆ 0, πœ€β€² β‰ˆ 0such that πœ€β€² πœ€ β‰ˆ 0,the inequalities (1)οΏ½οΏ½lim infπœ€β†’0𝐄𝛀��𝒙𝒕,,πœΊπ’™πŸŽ,πœΊβ€²(πœ”) βˆ’ 𝝀(𝟏)οΏ½ 𝟐 οΏ½οΏ½ πœ€β€²οΏ½ ≀ 𝐢̃ ′�𝑼�𝑑, 𝝀(𝟏)οΏ½οΏ½2, (2)οΏ½οΏ½lim infπœ€β†’0𝐄𝛀��𝒛𝒛𝑇,,𝜺𝟎,πœΊβ€²(πœ”) βˆ’ 𝝀(𝟐)οΏ½ 𝟐 οΏ½οΏ½ πœ€β€²οΏ½ ≀ 𝐢̃ ′‖𝑽(𝑇, 𝝀)β€–2, (99.b) (3)οΏ½οΏ½lim infπœ€β†’0π„πœ΄οΏ½οΏ½π’™π‘»,πœΊπ’™,πœΊβ€²(πœ”) βˆ’ π’šοΏ½ 𝟐 οΏ½οΏ½ πœ€β€²οΏ½ ≀ 𝐢̃ ′‖𝑼(𝑇, π’š)β€–2. (4) οΏ½π‰Μ…πœ€πœ€β€²,𝑗� πœ€β€² ≀ |𝑉𝑖(𝑇, π’š, 0)| + ‖𝑼(𝑇, π’š)β€– 2,𝑖 = 1, … , π‘š issatisfied.Here 𝒛𝒛𝑇,,𝜺𝟎,πœΊβ€²(πœ”) = ∫ π’ˆ0𝑇 πœΊβ€²οΏ½π’™π’™π‘»,𝜺𝟎,πœΊβ€²(πœ”), 𝜢(𝑑), 𝑑� 𝑑𝑑. Here a function𝑾(𝑑, 𝝀) = {𝑼(𝑑, 𝝀), 𝑽(𝑑, 𝝀)}𝐭 = οΏ½οΏ½π‘ˆ1(𝑑, 𝝀), … , π‘ˆπ‘›(𝑑, 𝝀)οΏ½; �𝑉1(𝑑, 𝝀), … , π‘‰π‘š(𝑑, 𝝀)��𝐭is the solution

(32)

𝑾̇(𝑑, 𝝀)= 𝐉�𝒃�0(𝝀, 𝑑)�𝑾(𝑑, 𝝀) + 𝒃�0(𝝀, 𝑑) + βŒ©π’…(𝑑), πœΆοΏ½π’(𝑑)βŒͺ,(100)

𝑼(0, 𝝀) = π’™πŸŽβˆ’ 𝝀, 𝑽(0, 𝝀) = 0.

And with the playoff of thei-th player is:

𝐉̆𝐒 = |𝑉𝑖(𝑇, π’š, 𝟎)| + ‖𝑼(𝑇, π’š)β€–2.(101)

Here

𝒃�0(𝝀, 𝑑) = {𝒇(𝝀, 0, 𝑑); π’ˆ(𝝀, 0, 𝑑)}t(102)

and

𝐉 = π‰οΏ½π’ƒοΏ½πŸŽ(𝝀, 𝑑)οΏ½(103)

isJacobian i.e., J is (𝑛 + π‘š) Γ— (𝑛 + π‘š)-matrix: π‰οΏ½π’ƒοΏ½πŸŽ(𝝀, 𝑑)οΏ½= JοΏ½πœ•π’ƒοΏ½πŸŽ,π’Š(𝒙, 𝑑)/πœ•π‘₯𝒋�𝒙=𝝀.(104)

Proof. Let us rewrite Eqs.(96)-(97) of the next equivalent form (i) �𝒙̇𝑑,πœ€π‘₯0,πœ€β€²(πœ”)οΏ½

πœ€β€² = οΏ½π’‡πœ€β€²οΏ½π’™π‘‘,πœ€β€²

π‘₯0,πœ€(πœ”), 𝜢(𝑑), 𝑑��

πœ€β€²+ +βˆšπœ€οΏ½π‘€(𝑑, πœ”)οΏ½πœ€β€² (105)

(ii) (ii) 𝒛̇𝑑,,πœΊπ’›0,πœΊβ€²(πœ”) = π’ˆπŸπœΊβ€²οΏ½π’™π‘»,πœΊπ’™0,πœΊβ€²(πœ”), 𝜢(𝑑), 𝑑� + +βˆšπœΊοΏ½π‘€(𝑑, πœ”)οΏ½πœΊβ€².(106)

Then the playoff of the i-th player is οΏ½π‰Μ…πœ€πœ€β€²,𝑖� πœ€β€²=𝐄 ��𝑧𝑇,,πœΊβ€²,π’Š 𝒛0,𝜺 (πœ”)οΏ½ πœ€β€²οΏ½ + 𝐄 οΏ½οΏ½βˆ‘ οΏ½π‘₯𝑇,πœ€β€²;𝑖 π‘₯,πœ€ (πœ”) βˆ’ 𝑦 𝑖� 2 𝑛 𝑖=1 οΏ½ πœ€β€²οΏ½. Here 𝑧𝑑,,πœΊπ’›0,πœΊβ€²,π’Š(πœ”) = ∫ 𝑔0𝑑 πœΊβ€²,π’ŠοΏ½π’™π’™π‘»,𝜺0,πœΊβ€²(πœ”), 𝜢(𝑑), 𝑑� 𝑑𝑑, 𝒛0 = 0.

The inequalities (99) immediately follow from Eq.(105), Eq.(106),Theorem 6and definitions.

(33)

with a small white noise.

(1)π‘₯Μ‡1 = π‘₯2, π‘₯Μ‡2 = βˆ’π‘˜π‘₯23+ 𝛼1(𝑑) + 𝛼2(𝑑) + βˆšπœ€π‘€(𝑑, πœ”); π‘˜ > 0, (107)

𝑑 ∈ [0, 𝑇], π‘₯1 (0) = π‘₯10, π‘₯2 (0) = π‘₯20 ; πœ€ β‰ͺ 1;

(2) 𝛼1(𝑑) ∈ [βˆ’πœŒ1, 𝜌1], 𝛼2(𝑑) ∈ [βˆ’πœŒ2, 𝜌2];

(3)𝐽𝑖 = π‘₯12(𝑇), 𝑖 = 1,2.

Optimal control problem for the first player is: min

𝛼1(𝑑) ∈ [βˆ’πœŒ1, 𝜌1] οΏ½

max 𝛼2(𝑑) ∈ [βˆ’πœŒ2, 𝜌2]

[π‘₯12(𝑇)]οΏ½(108)

and optimal control problem for the second player is: max

𝛼2(𝑑) ∈ [βˆ’πœŒ2, 𝜌2] οΏ½

min 𝛼1(𝑑) ∈ [βˆ’πœŒ1, 𝜌1]

[π‘₯12(𝑇)]οΏ½. (109)

Using Equation (100) one obtained the corresponding linear master game: (1)𝑒̇1 = 𝑒2, 𝑒̇2 = βˆ’3π‘˜πœ†22𝑒2βˆ’ π‘˜πœ†23 + 𝛼�1(𝑑) + 𝛼�2(𝑑), (110)

𝑒1 (0) = π‘₯10βˆ’ πœ†1, π‘₯2 (0) = π‘₯20βˆ’ πœ†2 ;

(2)𝛼�1(𝑑) ∈ [βˆ’πœŒ1, 𝜌1], 𝛼�2(𝑑) ∈ [βˆ’πœŒ2, 𝜌2];

(3)π‘±π’Š= 𝑒12(𝑇), 𝑖 = 1,2.

Optimal control problem for the first player is: min

𝛼�1(𝑑) ∈ [βˆ’πœŒ1, 𝜌1] οΏ½

max 𝛼�2(𝑑) ∈ [βˆ’πœŒ2, 𝜌2]

[𝑒12(𝑇)]οΏ½, (111)

References

Related documents

Sharar (2012) conducted a study about introducing and improving reflective practices of teachers in a school at Chitral, it was found that reflective practice is a big challenge

(Mordell-Weil). Neka je K polje algebarskih brojeva. Time je problem strukture torzijske grupe u potpunosti rijeˇsen te ostaje problem raˇcunanja r, broja nezavisnih

The Centre continued to develop and promote academic research in contribution to economic and social development of New Zealand and to public policy discussion and

β€’ Psoriatic skin has a fast mitotic rate?. β€’ Triggers an inflammatory response in and around

Keywords: discrete choice, decision making, risk, uncertainty, (cumulative) prospect theory,

like CSF leukocytes, proteins and glucose using urinary reagent strip and evaluated its utility and efficacy in the diagnosis of meningitis henceforth predicted that urinary

The Effect of Consumption of Unhealthy Snacks on Diet and the Risk of Metabolic Syndrome in Adults: Tehran Lipid and Glucose Study, Iran. Zahra

In this paper we propose the notion of Liquid Journals (LJ) as a way to overcome the information overload issue in scientific publications. Their underlying principles consist i)