The central theme of this doctoral thesis is understanding how solutions to certain nonlinear partial differential equations "blow up," meaning they develop singularities in finite time. When a solution blows up, mathematicians want to know exactly how the solution falls apart: does it concentrate energy into a small region, and if so, what does that concentrated piece look like? The thesis develops tools for decomposing blowing-up solutions into recognizable building blocks called profiles or bubbles, using a general framework built around the idea of "dislocation spaces" in Hilbert spaces, which provides a unified language for describing these decompositions across different equations.
One major contribution involves the nonlinear wave equation in four and five spatial dimensions, specifically the focusing, energy-critical version. The thesis constructs explicit solutions that blow up at a single point in finite time, with the blow-up rate controlled by a prescribed polynomial rule. The construction in four dimensions follows an established scheme that exploits the relatively clean algebraic structure of the equation. Five dimensions are significantly harder because the equation is less smooth, making it much more difficult to control the errors that accumulate during the construction. A modified technical setup is required to handle these rougher error terms, and this represents a genuine advance over prior methods.
The other main contribution concerns the harmonic map heat flow, which describes how a map from a disk to a sphere evolves over time to minimize a certain energy, analogous to how heat smooths out temperature distributions. The thesis proves that when such a flow develops a singularity in finite time, the solution concentrates into exactly one bubble, even under general boundary conditions, as long as the solution has a rotational symmetry called k-equivariance. The proof relies on comparison principles rather than the more common energy-based arguments. The thesis also shows that solutions can blow up in infinite time for any degree of this symmetry, and includes an appendix reviewing the classical treatment of this problem due to Struwe and Qing.