Tomoshi Otsuki

A SEARCH-BASED APPROACH TO RAILWAY ROLLING STOCK ALLOCATION PROBLEMS

Experts working for railway operators still have to devote much time and effort to creating plans for rolling stock allocation. In this paper, we formulate the railway rolling stock allocation problem as a set partitioning multi-commodity flow (SPMCF) problem and propose a search-based heuristic approach for SPMCF. We show that our approach can obtain an approximate solution near the optimum in shorter time than CPLEX for real-life problems. Since our approach deals with a wide variety of constraint expressions, it would be applicable to automatic development of practical plans for many railway operators.

A search-based approach to the railway rolling stock allocation problem

Experts working for railway operators still have to devote much time and effort to creating plans for rolling stock allocation. In this paper, we formulate the railway rolling stock allocation problem as a set partitioning multi-commodity flow (SPMCF) problem and we propose a search-based heuristic approach for SPMCF. We show that our approach can obtain an approximate solution near the optimum in shorter time than CPLEX for real-life problems. Since our approach deals with a wide variety of constraint expressions, it would be applicable for developing practical plans automatically for many railway operators.

Demand Response Optimization Based on Building’s Characteristics

Demand response (DR) is one of the technologies that targets for power control based on the cooperation between power suppliers and consumers. In the case of buildings’ power control, it is important for buildings to achieve buildings’ power reduction target more correctly under individual buildings’ less burden. We suppose the framework of buildings’ aggregator that collects the information for power reduction (NEGAWATT information) and requests each building to reduce demand to achieve buildings’ power reduction target efficiently. In the framework, we first collect the NEGAWATT information based on buildings’ characteristics, and then we make the demand response plans (DR plans) that meet the requirements of buildings. In this paper, we focus on the DR optimization techniques based on NEGAWATT information, and show two simulation results, one of which shows the aggregation effect in a simple simulation, the other of which shows the multiple scenario methods to deal with the DR optimization under uncertainty. These simulations show that DR utilizing NEGAWATT information is more efficient for demand-supply balance than conventional DR methods.