The purpose of this paper is to propose a new algorithm for obtaining approximate solutions to the Burgers' equation (BE). Integration in time by a quadratic B-spline collocation method is shown. To the best of our knowledge, B-splines have not previously been used to integrate partial differential equations in both time and space. First, the BE is integrated using quadratic B-spline functions in time, and then the time-integrated BE is further solved in space via the cubic B-spline collocation method. The resulting recursive algebraic equation is used to obtain both shock wave and front propagation solutions of the BE, demonstrating the effectiveness of the space--time B-spline collocation method.
- Paper ID: 2510.09408
- Title: A time-space B-spline integrator for the Burgers' equation
- Authors: Idris Dag, Serkan Uğurluoğlu, Nihat Adar (Eskisehir Osmangazi University, Turkey)
- Classification: math.NA cs.NA
- Publication Date: October 10, 2025 (arXiv preprint)
- Paper Link: https://arxiv.org/abs/2510.09408
This paper proposes a novel algorithm for solving approximate solutions to Burgers' equation. The method employs quadratic B-spline collocation for temporal integration. To the authors' knowledge, this is the first application of B-splines simultaneously in both temporal and spatial dimensions for integrating partial differential equations. The approach first integrates Burgers' equation in the temporal direction using quadratic B-spline functions, then further solves the time-integrated equation in the spatial direction through cubic B-spline collocation. The resulting recursive algebraic equations are utilized to obtain shock wave and traveling wave front solutions of Burgers' equation, demonstrating the effectiveness of the space-time B-spline collocation method.
- Significance of Burgers' Equation: Burgers' equation is a quasi-linear parabolic partial differential equation widely applied in engineering sciences for modeling phenomena such as turbulence, gas dynamics, heat conduction, elasticity, traffic flow, and acoustic waves. It provides a mathematical formulation describing the interaction between convection and diffusion.
- Necessity of Numerical Solutions: Nonlinear partial differential equations typically lack analytical solutions, particularly when the viscosity constant υ is small, resulting in slowly converging series solutions with limited practical utility. Small υ values lead to shock formation, making both analytical and numerical methods challenging to provide reasonable solutions.
- Limitations of Existing Methods: Traditional time integrators including Crank-Nicolson schemes, Runge-Kutta methods, geometric integrators, and exponential integrators face challenges in accuracy and stability when addressing shock waves and traveling wave fronts in Burgers' equation.
The core motivation of this paper is to develop a novel numerical method capable of:
- Simultaneously employing B-splines for integration in both temporal and spatial dimensions
- Effectively handling shock waves under small viscosity values
- Providing higher-precision numerical solutions
- Novel Methodology: First proposal of using B-splines simultaneously as temporal and spatial integrators for solving Burgers' equation
- Dual B-spline Strategy: Innovative combination of quadratic B-splines in the temporal direction and cubic B-splines in the spatial direction
- Shock and Traveling Wave Simulation: Successful simulation of shock wave solutions and traveling wave front solutions of Burgers' equation
- Recursive Algorithm Framework: Establishment of a recursive algebraic equation solving system based on B-spline collocation
Solving the one-dimensional time-dependent Burgers' equation:
ut+uux−νuxx=0,x∈[a,b],t∈(0,T]
with initial condition: u(x,0)=f(x)
and boundary conditions: u(a,t)=u0,ux(a,t)=u0′,u(b,t)=um,ux(b,t)=um′
where ν is the viscosity constant.
- Grid Partitioning: Divide the temporal domain [c,d] into grid points c=t0<t1<⋯<tk=d, where tj=t0+jΔt
- Quadratic B-spline Basis Functions: Define Bj(t)∈C1[a,b], j=−1,…,n:
(t_{j+2}-t)^2 - 3(t_{j+1}-t)^2 + 3(t_j-t)^2, & t_{j-1} \leq t \leq t_j \\
(t_{j+2}-t)^2 - 3(t_{j+1}-t)^2, & t_j \leq t \leq t_{j+1} \\
(t_{j+2}-t)^2, & t_{j+1} \leq t \leq t_{j+2} \\
0, & \text{otherwise}
\end{cases}$$
3. **Approximate Solution Representation**:
$$U(x,t) = \sum_{j=-1}^n \delta(x,t_j)B_j(t)$$
#### Spatial Discretization (Cubic B-splines)
1. **Spatial Grid**: Partition the spatial domain $[a,b]$ uniformly, $x_i = x_0 + ih$, $i = 0,\ldots,m$
2. **Cubic B-spline Basis Functions**: $B_i(x) \in C^2[a,b]$, $i = -1,\ldots,m$
3. **Parameter Expansion**:
$$\delta(x,t^j) = \sum_{i=-1}^{m+1} \sigma_i^j B_i(x)$$
### Technical Innovations
1. **Hierarchical Integration Strategy**: First integrate in the temporal direction using quadratic B-splines, then solve in the spatial direction using cubic B-splines
2. **Collocation Point Method**: Enforce the differential equation at grid points
3. **Linearization Treatment**: Linearize nonlinear terms by selecting values from the previous time step
4. **Iterative Refinement**: Employ iterative processes to enhance accuracy at each time step
## Experimental Setup
### Test Problems
#### Problem 1: Shock Wave Solution
Analytical solution:
$$u(x,t) = \frac{x/t}{1 + \sqrt{t/t_0}\exp(x^2/(4\nu t))}, \quad t \geq 1, \quad 0 \leq x \leq 1$$
where $t_0 = \exp(1/(8\nu))$
#### Problem 2: Traveling Wave Front Solution
Analytical solution:
$$u(x,t) = \frac{\alpha + \mu + (\mu - \alpha)\exp(\eta)}{1 + \exp(\eta)}$$
where $\eta = \frac{\alpha(x - \mu t - \gamma)}{\nu}$, $\alpha = 0.4$, $\mu = 0.6$, $\gamma = 0.125$
### Evaluation Metrics
- **$L_\infty$ norm**: $|U - U_N|_\infty = \max|U_i^j - (U_N)_i^j|$
- **$L_2$ norm**: $|U - U_N|_2 = \sqrt{\sum_i |U_i^j - (U_N)_i^j|^2}$
- **Peak value**: Maximum value of the shock wave
### Implementation Details
- Space-time step sizes: $h = \Delta t = 0.01, 0.005, 0.001$
- Viscosity parameters: $\nu = 0.01, 0.005$
- Time range: $t = 3.25$ (shock problem), $t = 1.2$ (traveling wave problem)
## Experimental Results
### Main Results
#### Shock Wave Problem Results
| $\nu=0.01$ | $h=\Delta t$ | $L_\infty$ | $L_2$ | Peak value |
|-------------|--------------|------------|-------|------------|
| | 0.01 | 0.0498 | 0.0211| 0.2233 |
| | 0.005 | 0.0508 | 0.0207| 0.2225 |
| | 0.001 | 0.0516 | 0.0206| 0.2203 |
| $\nu=0.005$ | $L_\infty$ | $L_2$ | Peak value |
|--------------|------------|-------|------------|
| | 0.0582 | 0.0166| 0.2503 |
| | 0.0571 | 0.0159| 0.2480 |
| | 0.0560 | 0.0155| 0.2467 |
#### Traveling Wave Problem Results
| $\nu=0.01$ | $h=\Delta t$ | $L_\infty$ | $L_2$ |
|-------------|--------------|------------|-------|
| | 0.01 | 0.4875 | 0.1138|
| | 0.005 | 0.2964 | 0.0638|
| | 0.001 | 0.1514 | 0.0354|
| $\nu=0.005$ | $L_\infty$ | $L_2$ |
|--------------|------------|-------|
| | 0.8841 | 0.2532|
| | 0.6906 | 0.1329|
| | 0.2427 | 0.0367|
### Experimental Findings
1. **Precision-Grid Relationship**: Employing smaller space-time step sizes yields smoother solutions and improved accuracy
2. **Viscosity Effects**: Smaller viscosity constants result in sharper wave forms and higher errors
3. **Error Distribution**: Maximum errors typically occur near shock peaks or wave front regions
4. **Stability**: The method maintains good numerical stability across different parameter settings
## Related Work
Major research directions in this field include:
1. **Traditional Numerical Methods**: Finite difference, finite element, and spectral methods
2. **B-spline Methods**: Primarily used for spatial discretization, rarely for temporal integration
3. **Collocation Methods**: Galerkin methods, collocation methods, and meshless methods
4. **Specialized Solvers**: Numerical methods specifically designed for Burgers' equation
The innovation of this paper lies in the first simultaneous application of B-splines to both temporal and spatial integration, filling a gap in the field.
## Conclusions and Discussion
### Main Conclusions
1. The proposed space-time B-spline collocation method effectively solves Burgers' equation
2. The method demonstrates good accuracy and stability for both shock wave and traveling wave front problems
3. Solution precision and smoothness can be controlled by adjusting space-time step sizes
### Limitations
1. **Computational Complexity**: Requires solving linear equation systems at each time step
2. **Nonlinear Treatment**: Necessitates linearization and iterative processes
3. **Parameter Sensitivity**: Relatively sensitive to viscosity and grid parameters
4. **Scope of Application**: Primarily focused on one-dimensional Burgers' equation
### Future Directions
The authors propose that future work will focus on:
- Using higher-order B-splines to enhance accuracy
- Extending to multidimensional Burgers' equations
- Optimizing algorithm efficiency
## In-Depth Evaluation
### Strengths
1. **Method Innovation**: First proposal of space-time B-spline integration method with originality
2. **Theoretical Completeness**: Relatively comprehensive from mathematical derivation to numerical implementation
3. **Experimental Sufficiency**: Tests two typical problem classes—shock waves and traveling wave fronts
4. **Accuracy Analysis**: Provides detailed error analysis and convergence studies
### Weaknesses
1. **Insufficient Theoretical Analysis**: Lacks rigorous theoretical proofs of convergence and stability
2. **Limited Comparative Experiments**: Insufficient comparison with other advanced methods
3. **Missing Computational Efficiency Analysis**: No provision of computation time and memory usage analysis
4. **Insufficient Parameter Selection Guidance**: Lacks theoretical guidance for B-spline order selection
### Impact
1. **Academic Contribution**: Provides new perspectives for PDE numerical solutions
2. **Practical Value**: Simple and easy to implement, suitable for engineering applications
3. **Extensibility**: Framework generalizable to other nonlinear PDEs
### Applicable Scenarios
This method is particularly suitable for:
- Scientific computing requiring high-precision solutions to Burgers' equation
- Numerical simulation of shock wave and traveling wave front phenomena
- Applications with high accuracy requirements but moderate efficiency demands
## References
The paper cites 16 relevant references covering numerical methods for Burgers' equation, B-spline theory, collocation methods, and other related fields, providing a solid theoretical foundation for this research.