There has been a lot of work that has been done on timing predictability of real-time tasks on embedded systems. The main assumption in these studies has been that the timing behavior has been based on single processor systems. The scenario has changed entirely when the single core systems have been replaced with the new Multicore systems. The timing predictability is controlled by the migrating tasks, the network topology connecting the cores and the number of cores on the system. In this thesis we come up with a feasibility analysis which depends on the characteristics of the tasks viz. number of cache lines, time of migration, available bandwith, number of tasks etc. We also test this analysis on novel mechanisms of migration which have been proposed recently and present its results.

Over the past few years, Deep Learning (DL), a promising technology leading the next generation of intelligent environments, has attracted significant attention and has been intensively utilized in various fields in the fourth industrial revolution era. The applications of Deep Learning in the area of Mobile Edge Computing (MEC) have achieved remarkable outcomes. Among several functionalities of MEC, the service migration frameworks have been proposed to overcome the shortcomings of the traditional methodologies in supporting high-mobility users with real-time responses. The service migration in MEC is a complex optimization problem that considers several dynamic environmental factors to make an optimal decision on whether, when, and where to migrate. In line with the trend, various service migration frameworks based on a variety of optimization algorithms have been proposed to overcome the limitations of the traditional methodologies. However, it is required to devise a more sophisticated and realistic model by solving the computational complexity and improving the inefficiency of existing frameworks. Therefore, an efficient service migration mechanism that is able to capture the environmental variables comprehensively is required. In this thesis, we propose an enhanced service migration model to address user proximity issues. We first introduce innovative service migration models for single-user and multi-user to overcome the users’ proximity issue while enforcing the service execution efficiency. Secondly, We formulate the service migration process as a complicated optimization problem and utilize Deep Reinforcement Learning (DRL) to estimate the optimal policy to minimize the migration cost, transaction cost, and consumed energy jointly. Lastly, we compare the proposed models with existing migration methodologies through analytical simulations from various aspects. The numerical results demonstrate that the proposed models can estimate the optimal policy despite the computational complexity caused by the dynamic environment and high-mobility users.

Some of the most important work in the development of economic theory is associated with the study of market structure. In essence, most markets are two-sided. For example, product markets connect tens of thousands of product brands to tens of millions of consumers; marriage markets couple the single men and women who would otherwise suffer from a lonely heart; and labour markets link the job candidates to their preferred employers and positions. Apart from the two-sidedness, we have explored another important common aspect of these market structures, i.e. interconnection/competition of the segments within one side of the market. Under this common thread, the three essays in this thesis are freshly formulated in a loosely related manner, covering topics in three different areas. Chapter 2 is motivated by strategic transitions of many marketplaces (e.g. Amazon.com). From the perspective of a platform owner, when it owns part of the business on one side of the market, there is no straightforward answer as to whether having the rest of business owned by others is advantageous or not. The argument is that, on the one hand, the platform welcomes more third-party business as it boosts revenue in terms of membership fees; on the other hand the business owned by the platform dislikes the incoming competitors whose participation drives down pro t margins. We propose a novel framework in this chapter to explore the trade-off between the two. Here, the intermediary can decide to be either a "merchant" or a "two-sided platform", or a hybrid one in between. Our analysis shows that in hybrid mode the platform extracts all the surplus from the producers of the merchandised brands, and the merchandised brands always charge a price premium compared to the directly retailed ones. We also show that as the platform absorbs an existing directly retailed brand into the self-brand portfolio, the equilibrium prices of both brand types are increased. We find that only the directly retailed brands dominate the market when the platform s capacity is relatively small; and both brand types coexist in the marketplace when the capacity is relatively large. Furthermore, we find a backward bending proportion plus a vertical proportion of the "contract curve" in comparative statics. That is, the self-brand portfolio always expands while the third-party-brand portfolio shrinks until it reaches a certain level, when the platform increases its capacity. It helps us to gain some ideas on the dynamics of brand portfolio management for the platform. Lastly, taking into account of indirect network effect which is the common feature in the two-sided market, it is shown that the platform is better o¤ when consumers have positive expected surplus. Chapter 3 is much motivated by the Chinese experience. China has witnessed the largest rural to urban labour ow (among which the majority are male) in the world s history over the last three decades. We propose an idea that the grand migration can also be attributed to the unbalanced sex ratio between rural and urban areas. This chapter develops a two-sided matching model of two linked marriage markets with homogeneous agents, non-transferable utility and search friction. We extend the one-market model of the previous literature into a two-market one, allowing the agents to migrate between the markets at a fixed cost. The analysis focuses on the unmatched as well as the migrating population, which is induced by the different sex ratios in the two geographically isolated marriage markets. We find that imperfections in the matching technology leads to the enlarged gap of sex ratio of the unmatched population compared to that of the unbalanced inflows. We are interested in the question of how the migrating costs affect the migration between rural and urban areas, and under what conditions a subsidy covering migrating costs might benefit a party in the marriage markets. We characterise the equilibrium set in the parameter space of migrating costs, and find that a full subsidy of migrating costs does not necessarily benefit those who receive it but always benefits the opposite sex, if they are the short sides of both markets. Chapter 4 explains the migration of labour force from a different angle. Here, the migration is of workers to jobs. Motivated by the distinction of public and private sector, we consider a spatial oligopsony model in which forms (two co-locating small firms with recruiting capacity constraints and a large firm without such limit) are competing for workers along a "strip" market. The capacity issue that is extensively discussed in the Chapter 2 again plays an important role in this model, though in a very different context. It is shown that the recruiting capacity affects the intra-group competition and hence the inter-group competition in wage- posting strategies. Additionally, we show that, as recruiting limits expand, the expected wages offered by the small firms increase while the wage offered by the big firm decreases, which helps to explain the recent trend of the wage disparity between public and private jobs. We also characterise the equilibrium wages and the size (direction) of the migration in the three-stage game (i.e. the workers decide whether to relocate in the first stage, then the big firm decides its wage offer, and lastly, the two co-locating firms simultaneous set wages), which helps us to understand better the inter-sector mobility in a changing environment of economy. We investigate the issues of interconnection and competition in three different markets. It is always of interest for a researcher of economics to have some ideas on the same issue from different perspectives. Remember that whilst this is a collection of essays on economic theory, it is nonetheless compared to empirical observation. And it will surely serve as a starting point for the author to further the research on market structure.

Le réchauffement climatique entraîne des changements dans la dynamique et la distribution des populations. J'ai utilisé une étude de 19 ans, en Bretagne, sur un limicole longévif et migrateur partiel, l'Avocette élégante, pour quantifier et comparer les paramètres démographiques associés aux différentes stratégies de migration. Les taux de survie et les manifestations de la sénescence associées étaient similaires chez les résidents et les migrateurs, mais les migrateurs montraient un âge de recrutement plus tardif que les résidents. L'investissement reproductif était plus élevé et exempt de sénescence chez les individus recrutés à l'âge d'un an, alors que ceux commençant à se reproduire plus tard subissaient de la sénescence reproductive. La fitness des migrateurs était inférieure à celle des résidents, ce qui explique leur déclin pendant la période étudiée, alors que la population résidente est elle restée stable. La faible productivité, due à la prédation, entraîne le déclin de la population bretonne malgré un taux d'immigration important, ce qui doit absolument être pris en compte lors de la définition des politiques locales de conservation. / Global warming causes changes in the dynamics and distribution of populations. I used a 19-year study, in Brittany, on a long-lived and partial migrant, the Pied Avocet, to quantify and compare the demographic rates associated with different migration strategies. Survival rates and associated senescence patterns were similar in residents and in migrants, but migrants exhibited a delayed recruitment age. Reproductive investment was higher and senescence was absent in individuals recruited at the age of one year, whereas those who began to reproduce later showed reproductive senescence. The fitness of migrants was lower than the fitness of residents, which explained their decline over the study period, while the resident population remained stable. Low productivity, due to predation, caused the Brittany population to decline despite a high immigration rate, which questions local conservation policies.