Oren E. Nahum

Data Mining and Machine Learning Tools for Combinatorial Material Science of All-Oxide Photovoltaic Cells

Growth in energy demands, coupled with the need for clean energy, are likely to make solar cells an important part of future energy resources. In particular, cells entirely made of metal oxides (MOs) have the potential to provide clean and affordable energy if their power conversion efficiencies are improved. Such improvements require the development of new MOs which could benefit from combining combinatorial material sciences for producing solar cells libraries with data mining tools to direct synthesis efforts. In this work we developed a data mining workflow and applied it to the analysis of two recently reported solar cell libraries based on Titanium and Copper oxides. Our results demonstrate that QSAR models with good prediction statistics for multiple solar cells properties could be developed and that these models highlight important factors affecting these properties in accord with experimental findings. The resulting models are therefore suitable for designing better solar cells.

Developing a model for the stochastic time-dependent vehicle-routing problem

Vehicle-routing problems (VRP) have been studied in depth. Many variants of the problem exist, most of them trying to find a set of routes with the shortest distance possible for a fleet of vehicles. This paper combines two important variants, the stochastic VRP and the time-dependent VRP, to form and define the Stochastic Time-Dependent VRP. An efficient heuristic that is a new variant of the well-known saving algorithm is introduced. The algorithm incorporates simulation that enables an estimate of each routes probability of being the quickest. This new algorithm yields fast results that are 10% higher than optimal solutions. Such results are similar to the performance of the saving algorithm when compared to the capacitated VRP.

A Framework for Solving Real-Time Multi-objective VRP

One of the most important logistics problems in the field of transportation and distribution is the Vehicle Routing Problem (VRP). In general, VRP is concerned with the determination of a minimum-cost set of routes for distribution and pickup of goods for a fleet of vehicles, while satisfying given constraints. Today, most VRPs are set up with a single objective function, minimizing costs, ignoring the fact that most problems encountered in logistics are multi-objective in nature (maximizing customers’ satisfaction and so on), and that for both deterministic and stochastic VRPs, the solution is based on a pre-determined set of routes. Technological advancements make it possible to operate vehicles using real-time information. Since VRP is a NP-Hard problem, it cannot be solved to optimality using conventional methods; therefore, the paper presents a heuristic framework for solving the problem. In real-time dynamic problems, a solution is given based on known data, as time progresses, new data are added to the problem, and the initial solution has to be re-evaluated in order to suit the new data. This is usually done at pre-defined time intervals. If the time intervals are small enough, thus, at each time interval the amount of information added is limited. Therefore, the new solution will be similar to the previous one. Due to the fact that the result is a solution set, not a single solution, and one solution is to be selected within a short time window, it is necessary to automatically select a single solution. For that, a framework, based on traditional and evolutionary multi-objective optimization algorithms, which incorporate multi-criteria decision making methods, for solving real-time multi-objective vehicle routing problems is presented.