Abstract:In this paper, we address a distributed control problem for a system of partial differential equations describing the evolution of a tumor that takes the biological mechanism of chemotaxis into account. The system describing the evolution is obtained as a nontrivial combination of a Cahn-Hilliard type system accounting for the segregation between tumor cells and healthy cells, with a Keller-Segel type equation accounting for the evolution of a nutrient species and modeling the chemotaxis phenomenon. First, we develop a robust mathematical background that allows us to analyze an associated optimal control problem. This analysis forced us to select a source term of logistic type in the nutrient equation and to restrict the analysis to the case of two space dimensions. Then, the existence of an optimal control and first-order necessary conditions for optimality are established.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to find the optimal control strategy in a system of partial differential equations that describes tumor growth. Specifically, this system combines the Cahn - Hilliard - type equation (used to describe the separation between tumor cells and healthy cells) and the Keller - Segel - type equation (used to describe the evolution of nutrients and chemotaxis phenomena). The goal of the paper is to optimize the process of tumor growth by selecting an appropriate control variable \(u\), thereby achieving effective control of tumor growth.
### Background of the Paper
- **Cahn - Hilliard Equation**: Used to describe the phase - separation phenomenon between tumor cells and healthy cells.
- **Keller - Segel Equation**: Used to describe the distribution of nutrients and the chemotaxis phenomena caused by them.
- **Logistic Source Term**: Introduced in the nutrient equation to prevent the solution from blowing up in a finite time.
### Research Problems
The paper mainly studies the following problems:
1. **Establishment of Mathematical Background**: Develop a robust mathematical framework for analyzing relevant optimal control problems.
2. **Existence of Optimal Control**: Prove that there exists an optimal control \(u^*\) such that the objective function \(J(u; \phi, \sigma)\) reaches its minimum value.
3. **First - Order Necessary Conditions**: Establish the first - order necessary conditions for optimal control, which are usually given in the form of variational inequalities.
### Objective Function
The objective function \(J(u; \phi, \sigma)\) is defined as:
\[
J(u; \phi, \sigma):=\frac{\alpha_1}{2}\int_Q|\phi - \varphi_Q|^2+\frac{\alpha_2}{2}\int_\Omega|\phi(T)-\varphi_\Omega|^2+\frac{\alpha_3}{2}\int_Q|\sigma - \varsigma_Q|^2+\frac{\alpha_4}{2}\int_\Omega|\sigma(T)-\varsigma_\Omega|^2+\frac{\alpha_5}{2}\int_Q|u|^2,
\]
where:
- \(\alpha_i\) (\(1\leq i\leq5\)) are non - negative constants, and not all of them are zero.
- \(\varphi_Q, \varphi_\Omega, \varsigma_Q, \varsigma_\Omega\) are given objective functions.
### Control Variable
The control variable \(u\) varies in an appropriate admissible control set \(U_{\text{ad}}\), which is defined as:
\[
U_{\text{ad}}:=\left\{u\in U:u_{\min}\leq u\leq u_{\max}\text{ a.e. in }Q\right\},
\]
where \(U := L^\infty(Q)\), and \(u_{\min}, u_{\max}\in L^\infty(Q)\) satisfy \(u_{\min}\leq u_{\max}\) almost everywhere.
### Main Results
- **Existence**: Proved the existence of the optimal control \(u^*\).
- **First - Order Necessary Conditions**: Gave the first - order necessary conditions for optimal control, which can be represented by the solution of an adjoint problem.
### Conclusion
Through rigorous mathematical analysis, the paper established the theoretical basis of optimal control for the system of partial differential equations that describes tumor growth, and proved the existence of optimal control and the first - order necessary conditions. These results provide theoretical support for tumor treatment in practical applications.