Journal articles on the topic 'Sliding Puzzle'

Today a number of everyday activities are done through the Internet. To perform such web services users must register in relation to websites or fill some form. In such websites, some hackers write malicious programs called bots that destroy website resources by creating fake registrations or form submissions. This false registration may adversely affect the performance of websites. Therefore, it is necessary to distinguish between actual human users and Web bots (or computer programs) via tests known as CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart). Most of the conventional CAPTCHA challenges have become an easy nut to crack for bots, hence calling a need for better CAPTCHA alternatives. This paper studies the infamous 'Sliding Puzzle' as an alternate CAPTCHA challenge and analysing the complexity to solve it via bot script.

The pathogenic consequences of 369 unique human HsMLH1 missense variants has been hampered by the lack of a detailed function in mismatch repair (MMR). Here single-molecule images show that HsMSH2-HsMSH6 provides a platform for HsMLH1-HsPMS2 to form a stable sliding clamp on mismatched DNA. The mechanics of sliding clamp progression solves a significant operational puzzle in MMR and provides explicit predictions for the distribution of clinically relevant HsMLH1 missense mutations.

One of the most common mechanisms used for speeding up problem solvers is macro-learning. Macros are sequences of basic operators acquired during problem solving. Macros are used by the problem solver as if they were basic operators. The major problem that macro-learning presents is the vast number of macros that are available for acquisition. Macros increase the branching factor of the search space and can severely degrade problem-solving efficiency. To make macro learning useful, a program must be selective in acquiring and utilizing macros. This paper describes a general method for selective acquisition of macros. Solvable training problems are generated in increasing order of difficulty. The only macros acquired are those that take the problem solver out of a local minimum to a better state. The utility of the method is demonstrated in several domains, including the domain of NxN sliding-tile puzzles. After learning on small puzzles, the system is able to efficiently solve puzzles of any size.

AbstractDNA mismatch repair (MMR) relies on MutS and MutL ATPases for mismatch recognition and strand-specific nuclease recruitment to remove mispaired bases in daughter strands. However, whether the MutS–MutL complex coordinates MMR by ATP-dependent sliding on DNA or protein–protein interactions between the mismatch and strand discrimination signal is ambiguous. Using functional MMR assays and systems preventing proteins from sliding, we show that sliding of human MutSα is required not for MMR initiation, but for final mismatch removal. MutSα recruits MutLα to form a mismatch-bound complex, which initiates MMR by nicking the daughter strand 5′ to the mismatch. Exonuclease 1 (Exo1) is then recruited to the nick and conducts 5′ → 3′ excision. ATP-dependent MutSα dissociation from the mismatch is necessary for Exo1 to remove the mispaired base when the excision reaches the mismatch. Therefore, our study has resolved a long-standing puzzle, and provided new insights into the mechanism of MMR initiation and mispair removal.

A pattern database (PDB) is a heuristic function implemented as a lookup table that stores the lengths of optimal solutions for subproblem instances. Standard PDBs have a distinct entry in the table for each subproblem instance. In this paper we investigate compressing PDBs by merging several entries into one, thereby allowing the use of PDBs that exceed available memory in their uncompressed form. We introduce a number of methods for determining which entries to merge and discuss their relative merits. These vary from domain-independent approaches that allow any set of entries in the PDB to be merged, to more intelligent methods that take into account the structure of the problem. The choice of the best compression method is based on domain-dependent attributes. We present experimental results on a number of combinatorial problems, including the four-peg Towers of Hanoi problem, the sliding-tile puzzles, and the Top-Spin puzzle. For the Towers of Hanoi, we show that the search time can be reduced by up to three orders of magnitude by using compressed PDBs compared to uncompressed PDBs of the same size. More modest improvements were observed for the other domains.

It is 50 years since the sliding of actin and myosin filaments was proposed as the basis of force generation and shortening in striated muscle. Although this is now generally accepted, the detailed molecular mechanism of how myosin uses adenosine triphosphate to generate force during its cyclic interaction with actin is only now being unravelled. New insights have come from the unconventional myosins, especially myosin V. Myosin V is kinetically tuned to allow movement on actin filaments as a single molecule, which has led to new kinetic, mechanical and structural data that have filled in missing pieces of the actomyosin–chemo–mechanical transduction puzzle.

To harness modern multicore processors, it is imperative to develop parallel versions of fundamental algorithms. In this paper, we compare different approaches to parallel best-first search in a shared-memory setting. We present a new method, PBNF, that uses abstraction to partition the state space and to detect duplicate states without requiring frequent locking. PBNF allows speculative expansions when necessary to keep threads busy. We identify and fix potential livelock conditions in our approach, proving its correctness using temporal logic. Our approach is general, allowing it to extend easily to suboptimal and anytime heuristic search. In an empirical comparison on STRIPS planning, grid pathfinding, and sliding tile puzzle problems using 8-core machines, we show that A*, weighted A* and Anytime weighted A* implemented using PBNF yield faster search than improved versions of previous parallel search proposals.

In real-time planning, the planner must select the next action within a fixed time bound. Because a complete plan may not have been found, the selected action might not lead to a goal and the agent may need to return to its current state. To preserve completeness, real-time search methods incorporate learning, in which heuristic values are updated. Previous work in real-time search has used table-based heuristics, in which the values of states are updated individually. In this paper, we explore the use of abstraction-based heuristics. By refining the abstraction on-line, we can update the values of multiple states, including ones the agent has not yet generated. We test this idea empirically using Cartesian abstractions in the Fast Downward planner. Results on various benchmarks, including the sliding tile puzzle and several IPC domains, indicate that the approach can improve performance compared to traditional heuristic updating. This work brings abstraction refinement, a powerful technique from offline planning, into the real-time setting.

The origin of instability or even disappearance of the superlubricity state in hydrogenated amorphous carbon (a-C:H) film in the presence of oxygen or water molecules is still controversial. Here, we address this puzzle regarding the tribochemical activities of sliding interfaces at the nanoscale. The results reveal that gaseous oxygen molecules disable the antifriction capacity of a-C:H by surface dehydrogenation of tribo-affected hydrocarbon bonds. In comparison, oxygen incorporation into the hydrocarbon matrix induces the formation of a low-density surface shear band, owing to which the friction state depends on the oxygen content. High friction of a-C:H film in humid environment originates from the “tumor-like” heterogeneous structures as formed in the highly oxidized tribolayer. Notably, an appropriate doping of silicon can completely shield the moisture effect by forming a silica-like tribolayer. These outcomes shed substantial lights upon the roadmap for achieving robust superlubricity of carbon films in a wide range of environments.

In real-time planning, an agent must select the next action to take within a fixed time bound. Many popular real-time heuristic search methods approach this by expanding nodes using time-limited A* and selecting the action leading toward the frontier node with the lowest f value. In this paper, we reconsider real-time planning as a problem of decision-making under uncertainty. We propose treating heuristic values as uncertain evidence and we explore several backup methods for aggregating this evidence. We then propose a novel lookahead strategy that expands nodes to minimize risk, the expected regret in case a non-optimal action is chosen. We evaluate these methods in a simple synthetic benchmark and the sliding tile puzzle and find that they outperform previous methods. This work illustrates how uncertainty can arise even when solving deterministic planning problems, due to the inherent ignorance of time-limited search algorithms about those portions of the state space that they have not computed, and how an agent can benefit from explicitly metareasoning about this uncertainty.

In real-time domains such as video games, planning happens concurrently with execution and the planning algorithm has a strictly bounded amount of time before it must return the next action for the agent to execute. We explore the use of real-time heuristic search in two benchmark domains inspired by video games. Unlike classic benchmarks such as grid pathfinding and the sliding tile puzzle, these new domains feature exogenous change and directed state space graphs. We consider the setting in which planning and acting are concurrent and we use the natural objective of minimizing goal achievement time. Using both the classic benchmarks and the new domains, we investigate several enhancements to a leading real-time search algorithm, LSS-LRTA*. We show experimentally that 1) it is better to plan after each action or to use a dynamically sized lookahead, 2) A*-based lookahead can cause undesirable actions to be selected, and 3) on-line de-biasing of the heuristic can lead to improved performance. We hope this work encourages future research on applying real-time search in dynamic domains.

Classic best-first heuristic search algorithms, like A*, record every unique state they encounter in RAM, making them infeasible for solving large problems. In this paper, we demonstrate how best-first search can be scaled to solve much larger problems by exploiting disk storage and parallel processing and, in some cases, slightly relaxing the strict best-first node expansion order. Some previous disk-based search algorithms abandon best-first search order in an attempt to increase efficiency. We present two case studies showing that A*, when augmented with Delayed Duplicate Detection, can actually be more efficient than these non-best-first search orders. First, we present a straightforward external variant of A*, called PEDAL, that slightly relaxes best-first order in order to be I/O efficient in both theory and practice, even on problems featuring real-valued node costs. Because it is easy to parallelize, PEDAL can be faster than in-memory IDA* even on domains with few duplicate states, such as the sliding-tile puzzle. Second, we present a variant of PEDAL, called PE2A*, that uses partial expansion to handle problems that have large branching factors. When tested on the problem of Multiple Sequence Alignment, PE2A* is the first algorithm capable of solving the entire Reference Set 1 of the standard BAliBASE benchmark using a biologically accurate cost function. This work shows that classic best-first algorithms like A* can be applied to large real-world problems. We also provide a detailed implementation guide with source code both for generic parallel disk-based best-first search and for Multiple Sequence Alignment with a biologically accurate cost function. Given its effectiveness as a general-purpose problem-solving method, we hope that this makes parallel and disk-based search accessible to a wider audience.

We explore a method for computing admissible heuristic evaluation functions for search problems. It utilizes pattern databases, which are precomputed tables of the exact cost of solving various subproblems of an existing problem. Unlike standard pattern database heuristics, however, we partition our problems into disjoint subproblems, so that the costs of solving the different subproblems can be added together without overestimating the cost of solving the original problem. Previously, we showed how to statically partition the sliding-tile puzzles into disjoint groups of tiles to compute an admissible heuristic, using the same partition for each state and problem instance. Here we extend the method and show that it applies to other domains as well. We also present another method for additive heuristics which we call dynamically partitioned pattern databases. Here we partition the problem into disjoint subproblems for each state of the search dynamically. We discuss the pros and cons of each of these methods and apply both methods to three different problem domains: the sliding-tile puzzles, the 4-peg Towers of Hanoi problem, and finding an optimal vertex cover of a graph. We find that in some problem domains, static partitioning is most effective, while in others dynamic partitioning is a better choice. In each of these problem domains, either statically partitioned or dynamically partitioned pattern database heuristics are the best known heuristics for the problem.

We describe how to convert the heuristic search algorithm A* into an anytime algorithm that finds a sequence of improved solutions and eventually converges to an optimal solution. The approach we adopt uses weighted heuristic search to find an approximate solution quickly, and then continues the weighted search to find improved solutions as well as to improve a bound on the suboptimality of the current solution. When the time available to solve a search problem is limited or uncertain, this creates an anytime heuristic search algorithm that allows a flexible tradeoff between search time and solution quality. We analyze the properties of the resulting Anytime A* algorithm, and consider its performance in three domains; sliding-tile puzzles, STRIPS planning, and multiple sequence alignment. To illustrate the generality of this approach, we also describe how to transform the memory-efficient search algorithm Recursive Best-First Search (RBFS) into an anytime algorithm.

Abstract The Saint-Gilles breccias, on the western flank of Piton des Neiges volcano, are clearly identified as debris avalanche deposits. A petrographic, textural and structural analysis of the breccias and inter-bedded autochthonous lava flows enables us to distinguish at least four successive flank slides. The oldest deposit sampled the hydrothermally-altered inner parts of the volcano, and has a large volume. Failure was favored by the presence of a deep intensely-weathered layer. The younger deposits are from superficial sources, as their products are rarely hydrothermalized and are more vesicular. The breccia formation, and especially the progressive breaking up occurring during the debris avalanche displacement, indicates the existence of high speed transport. In the Cap La Houssaye coastal area, abrasion and striation of the underlying lava formation, as well as the packing features observed in the breccia, are considered to be deceleration structures. Introduction Huge landslides of volcano flanks, whether or not initiated by magmatic intrusions, have been recognized as catastrophic events since the 1980 Mount St Helens eruption. On oceanic shield volcanoes, the contribution of failure to the edifice-building process was proposed by Moore [1964] and suggested elsewhere for Hawaii [Lipman et al., 1985 ; Moore et al., 1989], Reunion island [Lénat et al., 1989], Etna [McGuire et al., 1991], and Canarias [Carracedo, 1994, 1996 ; Marty et al., 1996]. This contribution is particularly obvious in island volcanoes showing a U-shaped caldera open to the ocean. Several mechanisms inherent to the causes of failure have been proposed, such as dyke intrusion [McGuire et al., 1990 ; Iverson, 1995 ; Voight and Elsworth, 1997], caldera collapse [Marty et al., 1997], or volcanic spreading [Borgia et al., 1992 ; van Wyk de Vries and Francis, 1997]. Invariably, other factors have been proposed as favorable to volcanic destabilization, such as the probable occurrence of deep low-cohesion layers due to the existence of pyroclastic or hyaloclastic layers [Duffield et al., 1982 ; Siebert, 1984] or an old basement. Gravity spreading models are now frequently proposed to explain the destruction of volcanic edifices [Borgia et al., 1992 ; Merle and Borgia, 1996 ; van Wyk de Vries and Borgia, 1996 ; van Wyk de Vries and Francis, 1997], most of them taking into account basal or intra-volcanic weakness zones. We propose that in such a scenario, density heterogeneity should be an important factor governing the slow evolution of the volcanic pile. Clague and Denlinger [1994] proposed a olivine-rich ductile basal layer that influences the stability of volcano flanks. On Reunion island, a large volcanic landslide has been proposed to explain the peculiar morphology of Piton de la Fournaise-Grand Brûlé [Vincent and Kieffer, 1978]. Bathymetric surveys [Bachèlery and Montagionni, 1983 ; Lénat et al., 1989, 1990 ; Cochonnat et al., 1990 ; Lénat and Labazuy, 1990 ; Labazuy, 1991 ; Bachèlery, 1995 ; Ollier et al., 1998] have confirmed the offshore occurrence of debris avalanche deposits. Similar deposits are also known to exist along the western, northern and southwestern submarine flanks of the Piton des Neiges volcano. Unlike other deposits showing inland prolongation, “Saint-Gilles breccias” displays a well-preserved and non-weathered texture and structure. Because of striking analogies between the “Saint-Gilles breccias” and, for example, the Cantal stratovolcano debris avalanche deposits [Cantagrel, 1995], we conclude that these formations are the products of repeated avalanches during the Piton des Neiges basaltic period [Bachèlery et al., 1996]. We propose an interpretation of their origin, emplacement mechanism and their role in the evolutionary process of the western flank of Piton des Neiges. Volcano-structural setting Mechanical instability of oceanic volcanic edifices generates huge flank landslides, with lateral and mainly submarine transport of sub-aerial materials. These landslides participate in the building of the lower submarine slopes of the volcano. Geophysical surveys have detected low cohesion materials in most offshore Reunion island areas [Malengrau et al., 1999 ; de Voogd et al., 1999 ; Lénat et al., 2001] showing that these materials have largely contributed to the construction of offshore Reunion Island. Such deposits are also found in the inner part (“Cirques”) of Piton des Neiges [Maillot, 1999]. On the other hand, electric and electromagnetic soundings have revealed a deep extending conductor within the Piton de la Fournaise volcanic pile [Courteaud et al., 1997 ; Lenat et al., 2000]. Interpretations about the nature and origin of this conductor depend on its location. In the central caldera zone, as revealed by SP positive anomalies [Malengrau et al., 1994 ; Zlotnicki et al., 1994], the hydrothermal and magmatic complex is probably responsible for the observed low resistivities. Along the flanks, such a hypothesis may not be realistic. Courteaud [1996] suggests the occurrence of a deep argilized layer of volcano-detritic origin. In any case, the hydrothermal complex with high fluid pressures and secondary minerals appears as a potential weak zone that may contribute to the volcano’s instability [Lopez and Williams, 1993 ; Frank, 1995]. Chronology and stratigraphy Extent of the debris avalanche deposits The various breccias found at the western end of Reunion island, on the Piton des Neiges volcano flank, cover a 16 km2 area between Cap Marianne and Saint-Gilles (fig. 1). They are overlain upwards (> 250 to 300 m) by trachyandesitic (mugearite) lava flows of Piton des Neiges differentiated series [Billard, 1974]. Some restricted breccia outcrops in deep valleys from Bernica to the north up to l’Hermitage to the south indicate the existence of larger extension of the debris avalanche deposits. Furthermore, breccias with similar “Saint-Gilles” facies appear down the Maïdo cliff to Mafate “Cirque” at an altitude 1300 m, beneath 600 m of mugearite and some olivine basalt flows. Unpublished electromagnetic data (CSAMT soundings) confirm the inland continuity of the “Saint-Gilles breccias” up to the Maïdo along the Piton des Neiges western flank, hidden by mugearitic flows. Available bathymetric surveys offshore Saint Paul – Saint Gilles areas show the obvious underwater prolongation of “Saint-Gilles breccias” : a shallow depth (< 100 m) plateau followed by a slope with hummocky surface down to 2 500 m depth [Bachèlery et al., 1996 and fig. 2]. From this data, the total surface of “Saint-Gilles” debris avalanche deposits is estimated as more than 500 km2. Chronology A coastal cliff, from Ravine Bernica to Boucan Canot, provides the best outcrop of the northern part of “Saint-Gilles breccias”, with a clear inter-bedding of breccia units and lava formations (photo 1and fig. 3). – The lower breccia unit (Br I), of unknown thickness, has a remarkable friable aspect and a grayish color. – The first autochthonous lava formation (L1) consists in thin pahoehoe olivine basalt flows filling large valleys dug into “Br I”. The top of this formation is striated by the overlying “Br II” unit (photo 2). – Breccia unit “Br II” is interbedded between L1 and L2 olivine basalts. More compact and massive, “Br II” is characterized by a reddish matrix and dark blocks, with many curved fracture surfaces. – On “Br II” or directly on L1, picritic basalt flows L2 are found, filling narrow valleys. – Breccia unit “Br III” lies on “Br II” with a striking sheared contact plane visible along the main road (photo 3). It is a typical debris avalanche deposit with large imbricate blocks within a fine-grained beige matrix. – Once again, basaltic flows of lava formation L3 fill a valley dug into “Br III” near Petite Anse river. – Breccia unit “Br IV” rests on L3 at Petite Anse, but its contact with “Br III” elsewhere is not clear. The facies of this unit is very similar to the “Br III”. All the breccia units are covered by basaltic and trachyandesitic flows from the end of the Piton des Neiges basaltic series, and differentiated series. In the Saint-Gilles river, two formations are superposed : picritic basalts (L4) have flowed on the “Br IV” breccia unit, latter aphyric trachy-andesitic (mugearite) flows (L6) overlapped L4 and the breccia landforms, reaching in places the coastal area. To the north, at Plateau Caillou, plagioclase-phyric basalt flows (L5) are found between mugearite and breccias. Elsewhere on Piton des Neiges, such flows are symptomatic of the transition from the basaltic series to the differentiated series [Billard, 1974]. The occurrence of autochthonous basaltic formations L1 to L3, inter-bedded with “Saint-Gilles breccias”, enables us to distinguish at least four superposed breccia units. Although the emplacement age of the lower “Br I” is not known precisely, it is overlain and therefore older than Cap Marianne pahoehoe lavas (L1) dated at 0.452 Ma [Mc Dougall, 1971]. On the other hand, the upper breccia units are younger than the pahoehoe olivine basalt at Cap la Houssaye dated at 0,435 Ma but older than L5 plagioclasic basalts dated at 0.35 Ma. Geological description of the “breccia sequence” In the synthetic lithologic log (fig. 4) of the Saint-Gilles area, autochthonous lava formations are clearly broken into four separate breccia units. Lava formations. – L1 formation consists of numerous thin pahoehoe olivine-rich to aphyric basaltic flows. Both L2 and L3 formations are characterized by a few thicker (decametric) olivine (frequently picritic) basalt flows. Breccia units. – All breccia units display common characteristics such as the universal association of two facies (photo 4) : (i) a matrix – sandy to silty – facies containing a non-sorted mixture of non-stratified heterogeneous materials ranging from granular size to blocky elements, (ii) coherent large blocks and large pieces (‘block’ facies) of various lithology such as lava flow, scorias, pyroclastics or other breccias ; blocks displaying frequent “jigsaw” features. The lower breccia unit “Br 1” (fig. 4) has a more compact but very heterogeneous aspect, with a chaotic distribution of blocks in a less-developed matrix. This unit is characterized by a deep hydrothermal alteration with a lot of zeolites, chlorite, clays, calcite and oxides. The upper breccia units, “Br II” to “Br IV” (fig. 4) are less heterogeneous than “Br I” because their matrix facies are more voluminous and because the matrix clearly separates the bigger blocks. In both facies, a great diversity of fresh lithologic types such as picritic basalt, olivine-phyric basalt, plagioclase-phyric basalt and aphyric more or less vesicular basalts, gabbro, dunite are found, with no or only few slightly zeolitised blocks. Plurimetric to metric blocks are severely fractured, disintegrated into millimetric to decimetric angular pieces. The frequent polygenic aspect is due to block juxtaposition or imbrication. The abundant matrix is composed of crushed rocks and mineral elements, fine-grained (< mm), showing frequent fluidity and bedding marks (photo 5). The very heterogeneous composition of the matrix is confirmed at a microscopic scale. On the contrary, cores of blocks appear as jigsaw-puzzle-like monolithologic pieces of various basaltic rocks. At their edges, disintegration leads to progressive mixing with neighboring blocks that feed the matrix. Discussion Originality of “Saint-Gilles breccias” “Saint-Gilles breccias” constitute one of the few cases [see also Cantagrel et al., 1999] of debris avalanche deposit outcroppings on the sub-aerial part of an oceanic shield volcano. The main part of the deposit is suspected to be offshore. Their hummocky surface in delineating parallel ridges can be compared to the one described offshore the Grand Brûlé area, east of Piton de la Fournaise [Bachèlery et al., 1996]. “Saint-Gilles breccias” were deposited after several Piton des Neiges flank slide events that were separated by basaltic flows. Repeated debris avalanches have also been proposed to explain Piton de la Fournaise offshore deposits [Lenat et al., 1990 ; Labazuy, 1991]. The occurrence of autochthonous interbedded lava formations is essential to interpret the thick piling up of slide material along Reunion volcano flanks as deposits of repeated avalanches at the same place, instead of as being the products of a single huge event. Many structural and textural features noticed in the upper breccia units reveal crucial information on the emplacement mechanism of debris avalanches. For instance, brecciated blocks are typical of progressive break-up during transport processes. Blocks can simply be fractured, or they can be so severely disintegrated that stretching and mixing with other blocks and matrix formation are observed. The observation of such phenomena implies the existence of numerous percussive events between rocks, as well as internal vibrations in the debris avalanche and therefore the existence of high-speed transport. Lava formations L1 underlying upper breccia units are truncated and strongly striated in a seaward direction (photo 2), parallel to the breccia morphological ridges. In the same way, internal contact surfaces between upper breccia units are shear planes underlain by cataclastic layers and lenses (photo 3). Such structures are interpreted as due to drastic deceleration effects of avalanches reaching a topographic leveling out in the coastal area. This concords with the occurrence of sub-vertical contact areas between the blocks and the matrix. These injections of matrix between the blocks are generated bottom-up from the shear plane at the moment of the sudden deceleration of the avalanche. Other fracture planes that are in accordance with the morphology of ridges, are found in “Br III” unit (see fig. 5). They are interpreted as the result of packing effects. Origin of flank failures Although the source area of breccia formations has not yet been clearly identified, it has to be in the central part of Piton des Neiges as seen in the western cliff of “cirque de Mafate”. Furthermore, “Br I” deeply weathered materials evidently come from the hydrothermalized core of the volcano. Though the “Br I” thickness is not known, the volume involved may be considerable and a part of this volume must constitute the main body of Saint-Gilles offshore deposits. The upper breccias units “Br II” to “Br IV” display very similar textures and lithologies, with dominant non-altered basaltic rocks from the “Phase II” building stage of Piton des Neiges [Billard, 1974]. These units are very thin in the coastal area of Cap La Houssaye (see fig. 2) despite a proximal facies (meaning a deposit in the transport zone nearer than the main deposit zone). They obviously originate from shallow flank slides of restricted extent. We suggest that the upper Saint-Gilles deposits are due to repeated events that produced thin high-speed debris avalanches. Emplacement modalities The morphology of “Saint-Gilles breccias”, or submarine deposits offshore Grand Brûlé (east of Piton de la Fournaise volcano), are typical of sliding movements along shallow depth shear planes (several hundred meters up to two kilometers) within the volcanic pile. But several levels of decollement are suggested by seismic refraction and reflection profiles offshore La Reunion, the deepest corresponding to the top of the preexisting oceanic sediments [de Voogt et al., 1999]. Until now, in Reunion Island, only shallow failures affecting the upper parts of volcanic edifices, with deposits on the lower slopes, have been positively identified. Conditions that trigger giant flank landslides affecting oceanic shields remain poorly understood but we can reasonably speculate that weak hydrothermally-altered layers in the inner part of the volcano favor these gravity-driven processes related to repeated dike injections. The “Saint-Gilles breccia” sequence is considered as a multiphase lateral collapse structure whose first event (“Br I”) was apparently the most voluminous. The corresponding deposit displays frequent hydrothermally-altered material symptomatic of originating from the Piton des Neiges core. Within Piton des Neiges, the low cohesive weathered layer is quite extensive [Nativel, 1978 ; Rançon, 1982] possibly reaching down the volcano flanks [Courteaud et al., 1997]. The interpretative scheme that we propose (fig. 6) in our evaluation of the conditions for the emplacement of Saint-Gilles sequence, takes into account the existence of such a mechanical discontinuity within the volcanic pile. We propose that the massive landslide failure of the west flank of Piton des Neiges volcano that produced the “Br I” breccia, provided efficient channels for younger Piton des Neiges lavas to reach the western and southwestern coastline. Morphological features, as well as radiometric data [Mc Dougall, 1971 ; Gillot and Nativel, 1982] and magnetic surveys [Lénat et al., 2001], yield evidence for preferential accumulation of lava during the last 0.5 m.y. (corresponding mainly to the differentiated series) in this part of the volcano. The relative asymmetry of Piton des Neiges was acquired by rift migration in response to the first huge landslide that produced the “Br I” unit of “Saint-Gilles breccia”, in the manner described by Lipman et al. [1990] for Mauna Loa volcano in Hawaii. The later repetition of flank collapses is consistent with similar structures on other oceanic islands. Since the first lateral collapse, the Piton des Neiges edifice was probably characterized by the existence of an asymmetrical steeper western flank where the old zeolite-rich “Br I” deposits possibly act as a detachment surface for later successive landslides which may have occurred recurrently over a short time interval.

Design grammars enable the formal representation of a vocabulary and rules that describe how designs can be synthesized just as the grammar rules of a spoken language define how to formulate valid, i.e., grammatically correct, sentences. Design grammars have been successfully applied in numerous engineering disciplines and enable the automated synthesis of designs within a defined design language. Design grammar development, however, is challenging and lacks methodological support. In this paper, a novel method is presented that supports the development and application of design grammars using transition graphs. In these, nodes represent generated designs and edges represent grammar rules that transform one design into another. Rather than using a tree structure to represent the possible application of rules, transition graphs are automatically generated and used to help designers better understand the developed grammar. The grammar designer is given feedback on (a) the rules, and (b) rule application sequences. This feedback can be used to (a) improve the grammar, and (b) apply it more efficiently. Two case studies, a gearbox synthesis task and a sliding tile puzzle, demonstrate the method. The results show the feasibility of the method to support design grammar development and application.