Railway transportation of hazardous materials including Toxic Inhalation Hazard, is crucial to North American economy. Although railway companies have favorable safety records in moving hazardous materials shipments, the possibility of spectacular events resulting from multicars incidents, however low, does exist, and the consequence can be potentially catastrophic in multiple fatalities. The rail disaster in Lac-Mégantic, Quebec, resulted in 47 fatalities and around $1.5 billion damages in 2013, is an example of low-probability high-consequence event. In this dissertation we aim at the development of analytical approaches considering the risk associated with hazardous materials in railway transportation. We study three versions of trip plan problems in the presence of hazardous materials, denoted as hazardous materials trip plan problems. In the first part of this dissertation we incorporate the blocking and train makeup decisions into the hazardous materials trip plan generation process, while limiting the total population exposures and environmental damages below the given thresholds. In evaluating the risk, we use aggregate measures, i.e., population exposures and environmental damages. We propose a non-linear mixed integer programming formulation for the considered problem. The solution of the model is NP-hard. In order to solve realistic size problem instances, a heuristic method is proposed by decomposing the problem into freight-to-block and block-to-train assignment problems. We then investigate more realistic hazardous materials trip plan problems by relaxing some of the assumptions. In the second part of this dissertation we incorporate risk-spreading functions into trip plan generation process and train scheduling decisions. For each riskspreading function, we present a mathematical formulation and then we design a heuristic method to solve realistic size problem instances. We continue this study by introducing joint hazardous material trip plan and pricing problems. We also relax the assumption of the information of the customer requests are known in advance. Accordingly, we introduce different categories of customers with the definition of specific treatment for each of them including accept/reject basis and particular delivery and price regulations. In particular, we grouped customer requests into two classes as follows:​ (a) traditional customers, who sign long term contracts with the carrier, must be fulfilled by the carrier’s own services, and their delivery and price quotations are set in advance and not subject to change;​ and (b) irregular customers, who make request for a carload moves less frequently and on an irregular basis, maybe outsourced/rejected because of (1) lack of train capacities, (2) additional risk exceeds the given risk thresholds, or (3) service level requirements. We propose two-phase heuristic to solve the considered problem. In the first phase, we solve a deterministic transportation planning and train timetabling problem for the known demands in advance. In the second phase, an optimization-based problem is built and solved at the arrival of the new request. Eventually, the dissertation ends with conclusion and further research recommendations.