Abelian surfaces of GL2-type as Jacobians of curves

Josep Gonzalez,Jordi Guardia,Victor Rotger
DOI: https://doi.org/10.4064/aa116-3-3
2004-09-20
Abstract: We study the set of isomorphism classes of principal polarizations on abelian varieties of GL2-type. As applications of our results, we construct examples of curves C, C'/\Q of genus two which are nonisomorphic over \bar \Q and share isomorphic unpolarized modular Jacobian varieties over \Q ; we also show a method to obtain genus two curves over \Q whose Jacobian varieties are isomorphic to Weil's restriction of quadratic \Q-curves, and present examples.
Number Theory,Algebraic Geometry
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to handle the Coulomb interactions between charged particles efficiently and accurately in computer simulations. Specifically, the paper explores how to simulate charged - particle systems through local algorithms, thereby avoiding the high computational complexity problems in traditional methods. ### 1. **Background and Problem Description** When simulating charged systems on a computer, due to the long - range nature of Coulomb interactions, directly calculating the Coulomb potential energy \( U_c=\sum_{i < j}\frac{e_i e_j}{4\pi\epsilon_0 r_{ij}}\) between all pairs of particles requires \( O(N^2)\) computational effort, which is impractical in large - scale systems. Currently, most large - scale codes use the Fast Fourier Transform (FFT) method, interpolating charges onto a grid to reduce computational complexity, but this still has some limitations. ### 2. **Main Contributions of the Paper** The paper proposes several methods based on local dynamics that can effectively handle the Coulomb interactions between charged particles: - **Molecular Dynamics (MD)**: By introducing an auxiliary electric field and modified Maxwell's equations, a new local molecular dynamics algorithm is proposed, which can effectively simulate charged - particle systems while maintaining locality. - **Monte Carlo (MC)**: A local Monte Carlo algorithm based on an auxiliary electric field is developed, which can significantly improve computational efficiency and is easy to parallelize. ### 3. **Specific Problems and Solutions** #### a. **Efficient Simulation of Coulomb Interactions** The paper shows how to derive Coulomb's law by solving Maxwell's equations and constructs an efficient simulation algorithm on this basis. By introducing an auxiliary electric field \( E\), the paper proposes a new partition function \( Z(\rho)\) and demonstrates how to satisfy Gauss's law by locally updating the electric field and particle positions, thereby ensuring the correctness of Coulomb interactions. #### b. **Charge Conservation and Continuity Equation** The paper emphasizes the importance of charge conservation and points out that if the continuity equation \(\frac{\partial\rho}{\partial t}+\nabla\cdot J = 0\) is violated, Gauss's law will fail, which in turn will affect the effective generation of Coulomb interactions. Therefore, the paper proposes some new charge - conservation integrators to ensure charge conservation during discretization. #### c. **Reducing Grid Errors** To reduce the errors caused by interpolating charges onto a grid, the paper introduces a polynomial spline interpolation algorithm to "adjust" the errors of monomer and dimer potential energies. This enables local algorithms to achieve higher precision in molecular dynamics simulations. #### d. **Advantages of the Monte Carlo Algorithm** Compared with molecular dynamics, the Monte Carlo algorithm is more efficient in some cases. For example, at low volume fractions, the Monte Carlo algorithm can be 80 times faster than molecular dynamics. In addition, the paper also proposes a cluster Monte Carlo algorithm, which can re - introduce gaps in the electric field spectrum when the conductivity is zero, thereby further improving computational efficiency. ### 4. **Conclusion** In general, this paper successfully solves the problem of efficient simulation of Coulomb interactions in charged - particle systems by introducing local algorithms based on Maxwell's equations. These algorithms not only improve computational efficiency but also perform well in parallelization and are suitable for large - scale parallel computing environments.