June 2021 Summaries

The Vehicle Routing Problem (VRP) is a well-known optimization challenge in logistics, especially relevant for businesses facing dynamic customer demands, such as seasonal spikes in orders. It extends the Traveling Salesperson Problem by involving multiple vehicles that need to efficiently deliver products from a central depot to various locations. VRP is NP-hard, necessitating the use of heuristics or approximation algorithms to find feasible solutions. Variations include Static VRP, where all customer information is known beforehand, and Dynamic VRP, where parameters like request arrival times are unpredictable. More complex models, such as the Dynamic Vehicle Routing Problem with Time Windows (DVRPTW), introduce constraints like service opening hours, aiming to minimize both route length and reaction time. Effective VRP solutions often require modeling concrete data, with synthetic data being generated using processes like the Poisson distribution for request arrivals. The article discusses various approaches to solving VRP, including heuristic methods and batching strategies, which are crucial when dealing with streaming requests. It highlights the importance of adapting routes in real-time to ensure efficient deliveries and maintain customer satisfaction.

MAGE is an open-source repository designed to enhance Memgraph's capabilities by supporting various programming languages, and it has recently introduced full support for Rust, complementing the existing C/C++ and Python options. Rust is highlighted for its ability to balance performance and ease of development, offering a middle ground between the deep memory management required by C/C++ and the quick prototyping capabilities of Python. The implementation of Rust support, which took about 16 weeks to complete, involved designing a Rust API that interfaces with the underlying C API, despite challenges like optimizing performance to minimize value copies. Future plans for the Memgraph ecosystem include further development of the library code and the addition of new query modules, with an invitation for community contributions to both the library code and practical applications.

In the realm of computer science, the concept of influencers has emerged similarly to how social media figures gain prominence, with significant attention given to authors who produce highly regarded articles. The DBLP computer science bibliography serves as an extensive resource for identifying influential computer scientists through citation analysis. By employing the PageRank algorithm, which assesses influence based on citation frequency and the prominence of citing authors, a network of around half a million authors and nearly 12 million interactions is analyzed. This method, implemented using Memgraph's open-source MAGE project, reveals notable figures in the field, such as Prof. Jure Leskovec, a key contributor to network analysis. The process demonstrates the power of PageRank in evaluating author influence, highlighting both its effectiveness with static data and the challenges posed by dynamic data flows, while suggesting the potential of online algorithms for real-time analysis. The discussion underscores the value of graph analytics for understanding influence dynamics in academic communities and encourages exploration of other algorithms within the MAGE library.

Memgraph has developed a free 10-day email course designed to teach the basics of Cypher, an open-source declarative language used for interacting with graph databases. Cypher is currently undergoing a vendor-independent standardization process and offers an intuitive way to work with property graphs, making it ideal for those interested in exploring graph technology. The course covers a range of topics, including the basics of Cypher, filtering results, structuring data, handling cases and duplicates, working with lists and aggregating functions, advanced queries, modifying and entering data, and understanding constraints and indexing, using the Europe road network dataset on Memgraph Playground as a practical example. As one of Memgraph's initial projects of this nature, they encourage participants to provide feedback via their Discord server.

GQLAlchemy is a Python library designed for developers who work with graph databases, particularly Memgraph, enabling them to create and manipulate graph-like structures using Python. By installing GQLAlchemy, users can easily execute Cypher queries on Memgraph, creating and connecting nodes such as ingredients and products in a graph database. The library offers tools like a query builder for constructing and executing queries intuitively, and although it currently supports the openCypher standard, there are plans to extend its capabilities to include NetworkX support and accommodate additional Graph Query Languages. Developers are encouraged to experiment with GQLAlchemy, contribute to its development on GitHub, and apply it in their projects, while adhering to the principle of not mixing nodes and edges.