Optimal control for a fourth-order nonisothermal tumor growth model of Caginalp type

2026

Read Optimal control for a fourth-order nonisothermal tumor growth model of Caginalp type (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-24108.txt.

Introduction

Cavalleri, Colli, and Rocca study distributed optimal control for a nonisothermal Caginalp-type phase-field model of tumor growth under thermal therapy. The system couples a possibly viscous Cahn-Hilliard equation for healthy and tumor phases with heat balance and nutrient reaction-diffusion equations, including chemotaxis and active transport. Hyperthermia serves as the control variable. The authors introduce a tracking-type cost, prove optimal controls exist, differentiate the control-to-state map, and derive first-order conditions through a variational inequality in the adjoint variables.

Catalog record

YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE106416
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.24108

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Optimal control for a fourth-order nonisothermal tumor growth model of Caginalp type Abstract We study a distributed optimal control problem for a nonisothermal Caginalp-type phase-field model that describes tumour growth under thermal therapy. The PDE system couples a possibly viscous Cahn–Hilliard equation, governing the evolution of the healthy and tumor phases, with an equation for the heat balance, and a reaction-diffusion equation for the nutrient concentration. Chemotaxis and active transport effects are taken into account, and hyperthermia appears as a control variable. We introduce a suitable tracking-type cost functional and show the existence of optimal controls. Then, we analyse the differentiability of the control-to-state operator and establish necessary first-order conditions expressed through a variational inequality involving the adjoint state variables. Key words: tumor growth model, Cahn–Hilliard system, nonisothermal model, optimal control, adjoint system, necessary optimality conditions. AMS (MOS) Subject Classification: 35Q93, 49J20, 49K20, 35Q92, 92C50. Giulia Cavalleri 1 e-mail: [email protected] Pierluigi Colli 2,3 e-mail: [email protected] Elisabetta Rocca 2,3 e-mail: [email protected] 1 Weierstrass Institute, Anton-Wilhelm-Amo-Str 39, 10117 Berlin, Germany 2 Department of Mathematics “F. Casorati”, University of Pavia, 27100 Pavia, Italy 3 Research Associate at the IMATI – C.N.R. Pavia, via Ferrata 5, 27100 Pavia, Italy 1 Introduction In this paper, we study an optimal control problem for a nonisothermal tumor growth model, which was introduced and proved to be well-posed in our previous work [ 7 ] . Let T>0 be the final time and \Omega\subseteq\mathbb{R}^{d} be a sufficiently regular spatial domain in dimension d=2,3 , having outward unit normal vector on \partial\Omega given by \boldsymbol{\nu} . The domain \Omega represents the portion of the body under investigation. The main variables whose evolution is studied are: • the relative temperature \theta in \Omega ; • the phase-field variable \varphi , describing the difference in volume fraction between tumor and healthy cells (in particular, the regions \{\varphi=1\} and \{\varphi=-1\} correspond to tumor and healthy tissue, respectively), together with its associated chemical potential \mu ; • the concentration \sigma of a nutrient (e.g., oxygen or glucose) dissolved in the blood. For an assigned control u , representing a thermal therapy administered to the patient in the region under consideration, the corresponding state (\theta,\varphi,\mu,\sigma) is the unique solution of the PDE system \displaystyle\partial_{t}(\theta+\ell\varphi)-\Delta\theta=u, (1.1a) \displaystyle\partial_{t}\varphi-\Delta\mu=(\lambda_{P}\sigma-\lambda_{A}-\lambda_{E}\theta)\mathbbm{h}(\varphi), (1.1b) \displaystyle\mu=\tau\partial_{t}\varphi-\Delta\varphi+\beta(\varphi)+\pi(\varphi)-\chi\sigma-\Lambda\theta, (1.1c) \displaystyle\partial_{t}\sigma-\Delta(\sigma-\chi\varphi)=-\lambda_{C}\sigma\mathbbm{h}(\varphi)+\lambda_{B}(\sigma_{B}-\sigma)-\lambda_{D}\sigma\mathbbm{k}(\theta), (1.1d) posed in the parabolic cylinder Q\coloneqq\Omega\times(0,T) and supplemented with no-flux boundary conditions \partial_{\boldsymbol{\nu}}\theta=\partial_{\boldsymbol{\nu}}\varphi=\partial_{\boldsymbol{\nu}}\mu=\partial_{\boldsymbol{\nu}}\sigma=0 (1.2) on \Sigma\coloneqq\partial\Omega\times(0,T) , reflecting the idealised assumption that the portion \Omega of the body is isolated from its surroundings, and with the initial conditions \theta(0)=\theta_{0},\quad\varphi(0)=\varphi_{0},\quad\sigma(0)=\sigma_{0} (1.3) in \Omega . Here, the well-known (possibly viscous) Cahn–Hilliard reaction-diffusion subsystem ( 1.1b )–( 1.1d ) governs the evolution of the tumor phase parameter \varphi and the nutrient concentration \sigma , and is coupled with the temperature equation ( 1.1a ). In particular, the coupling between the relative temperature \theta and the tumor variable \varphi is of Caginalp type (cf. [ 5 ] ). For the reader’s convenience, we briefly recall the main features of the model, referring to the introduction of [ 7 ] for a more comprehensive discussion. The positive constants \ell and \Lambda are related to the latent heat of the tissue. We suppose that tumor cells proliferate by consumption of the nutrients and die because of apoptosis, with \lambda_{p} and \lambda_{a} being, respectively, the proliferation and apoptosis constant rates. Moreover, following [ 7 ] , we assume that \theta has cytotoxic effects proportional to the temperature, which we incorporate in the right-hand side of equation ( 1.1b ) through the term -\lambda_{E}\theta , where \lambda_{E} is a fixed positive parameter. The function \mathbbm{h} , multiplying the mass source term in equation ( 1.1b ), guarantees that the phenomena we have just described are proportional to the tumor cells available in a certain area. For example, \mathbbm{h} may be a monotone increasing function which is 0 where \{\varphi=-1\} and 1 where \{\varphi=1\} . The nonnegative constant \chi denotes a transport coefficient accounting for chemotaxis and active transport effects. The term \beta(\varphi)+\pi(\varphi) represents the derivative of a double-well potential. We allow for a fairly general polynomial growth, so that, in particular, the classical choice \beta(\varphi)+\pi(\varphi)=\varphi^{3}-\varphi is admissible. In contrast, singular potentials of logarithmic type as well as double-obstacle potentials fall outside the frame of the present analysis. The parameter \tau\geq 0 may be either strictly positive or zero. When \tau>0 , the term \tau\partial_{t}\varphi in ( 1.1c ) acts as a viscous regularisation in the phase-field dynamics. The case \tau=0 instead corresponds to the classical (non-viscous) Cahn–Hilliard system. Both regimes are addressed in our analysis. From a mathematical viewpoint, the presence of the viscous term improves the regularity of solutions, which may be advantageous in view of possible extensions to singular potentials. Concerning the right-hand side of ( 1.1d ), the term -\lambda_{C}\sigma\mathbbm{h}(\varphi) models nutrient consumption by tumor cells, with a higher consumption rate in regions where the tumor cell density is larger. The term \lambda_{B}(\sigma_{B}-\sigma) describes the supply of nutrients from the pre-existing vasculature. Finally, the function \mathbbm{k} accounts for the influence of temperature on nutrient uptake, for instance, through vasodilation effects. The control problem. Within this framework, a well-chosen heat source u could drive the system toward a tumor reduction or, more generally, toward a desired evolution or final state. We introduce some target functions \theta_{Q},\varphi_{Q} , and \theta_{\Omega},\varphi_{\Omega} representing, respectively, the desired evolution of the temperature and of the phase parameter, and the desired temperature and tumor proportion at the final time T . For example, the target \varphi_{\Omega} could be a suitable configuration for surgery. Then, we define the cost functional \begin{split}\mathcal{J}(\theta,\varphi,u)\coloneqq\ &\frac{b_{1}}{2}\int_{Q}|\theta-\theta_{Q}|^{2}\,\mathrm{d}x\,\mathrm{d}t+\frac{b_{2}}{2}\int_{\Omega}|\theta(T)-\theta_{\Omega}|^{2}\,\mathrm{d}x\\ &+\frac{b_{3}}{2}\int_{Q}|\varphi-\varphi_{Q}|^{2}\,\mathrm{d}x\,\mathrm{d}t+\frac{b_{4}}{2}\int_{\Omega}|\varphi(T)-\varphi_{\Omega}|^{2}\,\mathrm{d}x\\ &+\frac{b_{5}}{2}\int_{Q}|u|^{2}\,\mathrm{d}x\,\mathrm{d}t.\end{split} (1.4) The first four terms in ( 1.4 ) are of standard tracking-type, while the last regularising term prevents the administration of an excessive amount of therapy, thereby avoiding damage to other vital organs. Firstly, we aim at minimising the cost functional \mathcal{J} subject to the PDE system ( 1.1 )–( 1.3 ) and to the following admissible control set \mathcal{U}_{\text{ad}}\coloneqq\left\{u\in\mathcal{U}\,|\,u_{\textrm{min}}\leq u\leq u_{\textrm{max}}\text{ a.e. in }Q\right\}, (1.5) where \mathcal{U}\coloneqq L^{\infty}(Q) , and u_{\textrm{min}},u_{\textrm{max}} are assigned elements of \mathcal{U} satisfying u_{\textrm{min}}\leq u_{\textrm{max}} almost everywhere in Q . Notice that \mathcal{U}_{\text{ad}} is a nonempty, closed, bounded and convex subset of \mathcal{U} . In particular, there exists a constant M>0 such that \mathcal{U}_{\text{ad}}\subseteq\mathcal{U}_{M} , where the open set \mathcal{U}_{M} is given by \mathcal{U}_{M}\coloneqq\left\{u\in\mathcal{U}\,\middle|\,\lVert u\rVert_{L^{\infty}(Q)}<M\right\}. (1.6) Rephrasing the optimal control problem more compactly, it reads as follows: \min_{u\in\mathcal{U}_{\text{ad}}}\mathcal{J}(\theta,\varphi,u). (1.7) The subsequent, more challenging, goal is to characterise the optimal controls, i.e., the minimisers, through their necessary first-order optimality conditions. Related literature. Let us recall some of the literature related to tumor growth models based on Cahn–Hilliard type dynamics, which is expanding rapidly, reflecting the strong interest it has attracted from both pure and applied mathematicians. In most of the previous papers on the subject, the isothermal case was considered, assuming, indeed, constant temperature. The reader can refer, e.g., to [ 22 , 26 , 27 , 17 ] and reference therein, where the nutrient dynamic is taken into account by means of a reaction-diffusion equation analogous to our ( 1.1d ). Moreover, various works also incorporate the evolution of other relevant quantities like elastic effects, and fluid flows (see, e.g., [ 14 , 6 , 25

Read the complete text