Abstract. There has been a great deal of progress in the fifteen years that have elapsed since Wigderson published his survey on the complexity of the graph connectivity problem [Wig92]. Most significantly, Reingold solved the longstanding question of the complexity of the s-t connectivity problem in undirected graphs, showing that this is complete for logspace (L) [Rei05]. This survey talk will focus on some of the remaining open questions dealing with graph reachability problems. Particular attention will be paid to these topics: – Reachability in planar directed graphs (and more generally, in graphs of low genus) [ADR05,BTV07]. – Reachability in different classes of grid graphs [ABC + 06]. – Reachability in mangroves [AL98]. The problem of finding a path from one vertex to another in a graph is the first problem that was identified as being complete for a natural subclass of P; it was shown to be complete for nondeterministic logspace (NL) by Jones [Jon75]. Restricted versions of this problem were subsequently shown to be complete for other natural complexity

We study the complexity of restricted versions of s-t-connectivity, which is the standard complete problem for NL. In particular, we focus on different classes of planar graphs, of which grid graphs are an important special case. Our main results are: • Reachability in graphs of genus one is logspace-equivalent to reachability in grid graphs (and in particular it is logspace-equivalent to both reachability and non-reachability in planar graphs). • Many of the natural restrictions on grid-graph reachability (GGR) are equivalent under AC 0 reductions (for instance, undirected GGR, outdegree-one GGR, and indegree-one-outdegree-one GGR are all equivalent). These problems are all equivalent to the problem of determining whether a completed game position in HEX is a winning position, as well as to the problem of reachability in mazes studied by Blum and Kozen [BK78]. These problems provide natural examples of problems that are hard for NC 1 under AC 0 reductions but are not known to be hard for L; they thus give insight into the structure of L. • Reachability in layered planar graphs is logspace-equivalent to layered grid graph reachability (LGGR). We show that LGGR lies in UL (a subclass of NL). • Series-Parallel digraphs (on which reachability was shown to be decidable in logspace by Jakoby et al.) are a special case of single-source-single-sink planar directed acyclic graphs (DAGs); reachability for such graphs logspace reduces to single-source-single-sink acyclic grid graphs. We show that reachability on such grid graphs AC 0 reduces to undirected GGR. • We build on this to show that reachability for single-source multiple-sink planar DAGs is solvable in L. 1

We report progress on the NL vs UL problem.- We show unconditionally that the complexity class ReachFewL ⊆ UL. This improves on the earlier known upper bound ReachFewL ⊆ FewL.- We investigate the complexity of min-uniqueness- a central notion in studying the NL vs UL problem. – We show that min-uniqueness is necessary and sufficient for showing NL = UL. – We revisit the class OptL[log n] and show that ShortestPathLength- computing the length of the shortest path in a DAG, is complete for OptL[log n]. – We introduce UOptL[log n], an unambiguous version of OptL[log n], and show that (a) NL = UL if and only if OptL[log n] = UOptL[log n], (b) LogFew ≤ UOptL[log n] ≤ SPL.- We show that the reachability problem over graphs embedded on 3 pages is complete for NL. This contrasts with the reachability problem over graphs embedded on 2 pages which is logspace equivalent to the reachability problem in planar graphs and hence is in UL.

Reachability and distance computation are known to be NL-complete in general graphs, but within UL ∩ co-UL if the graphs are planar. However, finding longest paths is known to be NP-complete, even for planar graphs. We show that with the combination of planarity and acyclicity, finding the length of the longest path (and also enumerating one such path) is also in UL ∩ co-UL. The result extends to toroidal DAGs as well. We also address the question of when reachability, distance, and longest path are indeed equivalent on DAGs, and give partial bounds. When the number of distinct paths is bounded by a polynomial, counting the number of paths is known to be in NL. We show that for planar DAGs with this promise, counting can be done by a UAuxPDA in polynomial time. The UAuxPDA(poly) bound also holds if we want to compute the number of longest paths, or shortest paths, and this number is bounded by a polynomial (irrespective of the total number of paths). Along the way, we show that counting in general DAGs is possible in LogDCFL provided the number of paths is bounded by a polynomial and the target node is the only sink.

The Graph Isomorphism problem restricted to planar graphs has been known to be solvable in polynomial time many years ago. In terms of complexity classes however, the exact complexity of the problem has been established only very recently. It was proved in [6] that planar graph isomorphism can be computed within logarithmic space. Since there is a matching hardness result [12], this shows that the problem is complete for L. Although this could be considered as a result in algorithmics its proof relies on several important new developments in the area of logarithmic space complexity classes and reflects the close connections between algorithms and complexity theory. In this column we give an overview of this result mentioning the developments that led to it. 1

log-space algorithm for reachability in planar acyclic digraphs

In this dissertation, we present three research directions related to the question whether all randomized algorithms can be derandomized, i.e., simulated by deterministic algorithms with a small loss in efficiency. Typically-Correct Derandomization A recent line of research has considered “typically-correct” deterministic simulations of randomized algorithms, which are allowed to err on few inputs. Such derandomizations may be easier to obtain and/or be more efficient than full derandomizations that do not make mistakes. We further the study of typically-correct derandomization in two ways. First, we develop a generic approach for constructing typically-correct derandomizations based on seed-extending pseudorandom generators, which are pseudorandom generators that reveal their seed. We use our approach to obtain both conditional and unconditional typically-correct derandomization results in various algorithmic settings. For example, we present a typically-correct polynomial-time simulation for every language in BPP based on

We show that two complexity classes introduced about two decades ago are equal. ReachUL is the class of problems decided by nondeterministic log-space machines which on every input have at most one computation path from the start configuration to any other configuration. ReachFewL, a natural generalization of ReachUL, is the class of problems decided by nondeterministic log-space machines which on every input have at most polynomially many computation paths from the start configuration to any other configuration. We show that ReachFewL = ReachUL. 1

Abstract. We explore the restrictiveness of planarity on the complexity of computing the determinant and the permanent, and show that both problems remain as hard as in the general case, i.e. GapL and #P complete. On the other hand, both bipartite planarity and bimodal planarity bring the complexity of permanents down (but no further) to that of determinants. The permanent or the determinant modulo 2 is complete for ⊕L, and we show that parity of paths in a layered grid graph (which is bimodal planar) is also complete for this class. We also relate the complexity of grid graph reachability to that of testing existence/uniqueness of a perfect matching in a planar bipartite graph. 1